Fast, Cheap and Flashy An Indie Art Direction Adventure

[Music] hello everybody hi uh i'm adam degrandis this is my beard i'm sorry today uh i was the art director on a game that's about to be released called uh tooth and tail and i'm gonna be talking about a major art overhaul that happened in what we naively thought was uh maybe not like the 11th hour but like the 9th or 10th it turned out to be more like the second i'll talk about that chestnut later so this is going to be broken up into four parts the first part i'm going to be talking about the analysis of the old art how we kind of broke it down then i'm going to move on to kind of the initial corrective measures we took then i'm going to talk about production and then the fourth part is kind of the ramifications of all this uh and and the results of it uh so probably uh i'm not a big name in games a little bit of uh info about me before i start is probably helpful uh this is the picture i use across all my social media and it's jokey but uh it is a pretty genuine thing the the the small handful of people in here who really know me know that this is like yeah okay that's sure that's adam but that doesn't really tell you uh anything about my professional history just kind of where my headspace is so let's do some professional history i currently run a visual design in our production studio called chickadee with my wife sarah i'm also a part-time educator at maine college of art in portland maine um and i've been making games professionally since uh about o'fort now when i got started i didn't realize that people specialized i thought if you were going to make uh games if you're going to be an artist in games you had to be a concept artist and a technical artist and a visual effects artist and blo i didn't i didn't realize that there were distinctions between those things so and i'm self-taught so i took a very long time learning how to do all of that stuff and it was very very sweaty but it turns out that someone who understands that whole tapestry of game art is really really valuable to small teams uh small teams like pocket watch games who make tooth and tail and that brings us to tooth and tail so what is tooth and tail um tooth and tail is an arcade real-time strategy game that that tries to take the like the big broad pillars of the genre so uh you know um fielding armies and and building bases and all that kind of stuff uh and it cuts out all like the micro bloat that's that's crept into the genre since uh since it's hey day in the 90s um it's it's designed to be a strategy game as much for fans of the genre as like people who think they don't like strategy games right it's meant to be really approachable uh for a really long time we were using starcraft meets mario kart uh as the elevator pitch and i still think that's pretty apt so here's just a quick video this is just from a few weeks ago just to kind of ground everyone so the player plays a rat commander rallies various units to them it's designed to be played with the controller that's the end of the video and here's where it looked like when i joined the team now you may be thinking like oh yeah no big deal early artwork looks rough and that is true one of my very favorite one sentence lessons for my students is games look terrible until they look great but these shots are from about like 18 months into development they were considered full production this is what would ship so they're pretty rough in that regard uh the kicker was that the game played great it was really really fun and it was absolutely a delivery on the promise of accessible strategy but these rough visuals were kind of a barrier that stopped players and press from ever discovering that so in early 2015 andy schatz the kind of main pocket watch guy got in touch with me and said that the project needed a dedicated art person to kind of help right this ship but there were a couple really critical wrinkles a lot of money had already been spent on the project so this isn't a problem that could be solved with money and related to that the game needed to be shippable by the end of the year so about 10 months i've worked with pocket watch on and off for about 12 years and it's always a really really good time plus i myself am most happy trying to work through like a really meaty uh creative challenge and this was definitely one of those so i joined the team in february of 2015. so the very first thing i did was a big analysis of all that pre-existing art and i tried to articulate in objective terms uh why i thought it was coming up short here's some of that stuff art fundamentals so basically the stuff that you would learn in art class uh they weren't being used to communicate important information to the player uh so value the artists in the room know what that is it's how light and dark something is uh basically so it's great for making things stand out so what's the important stuff in an rts well that's going to be your units or your buildings or in the case of tooth and tail your commander you want those things to be higher contrast so darker darks and brighter brights all in the same area and you want the less critical stuff random terrain and whatever you want that to be a little bit more right in the middle a little bit more even if we if we look at one of these old images in grayscale we can kind of see that that that rat commander right in the middle is kind of those grays are kind of following uh falling right along the same level of gray as the rest of the image that that rats not really popping out because of value in the train itself there's these kind of like undulating darks and lights but there's no real rhyme or reason to it the one effective bit going on is those pigs which are important part of the game those are really popping off that dark uh terrain but overall this is just not a very effective use of value color wasn't really being used effectively either so we have one smart bit going on we've got this kind of orange red rat on top of mostly green terrain and that's a pretty classic example of complimentary colors uh which is similar to how the there's a joke about how all movie posters are blue and orange uh and it is kind of funny to see all of these spread out they all kind of look the same but the fact the matter is complementary colors can be pretty striking when used skillfully but in the case of tooth and tail it wasn't really pushed far enough and it wasn't consistent for instance if i played as any other commander my faction color would change and i would lose that complementary color relationship additionally speaking of faction color if we look at the faction color of the units it's red but those reds are really dark and kind of under saturated the saturation is lower than the terrain so they don't really pop off the terrain overall the color palette kind of felt a little subdued and sleepy and definitely didn't prime the player for the kind of like fast uh quick action that the the game is kind of made of um related to those art fundamentals character designs were overly complex now i'm a big believer that all design um character design game design industrial design graphic design all of it it's not really about what you can add but what you can take away the units in tooth and tail when i joined the team were like detailed little paintings which made them really really cool paintings but kind of poor character designs when you got a bunch of them in a mob which is a pretty common occurrence in the game uh all that detail kind of just became this undulating block of noise it made it really hard to see what was actually going on uh one of the biggest issues here was that silhouette wasn't consciously considered when designing the units all the units had kind of a similar posture a similar volume a lot of them carried kind of similarly sized weapons now this maybe seems insignificant or even silly talking about silhouette but it's actually one of the most powerful tools in a character designers toolkit it helps maybe more than anything that five second read something to kind of think about it's not an accident that there are no background characters in the simpsons that have the same like head shape as lisa or bart or marge if you want something to stand out in a crowd give it a unique silhouette and and even better put some useful information into that silhouette another problem the visual design of the environment was straddling between organic and board game like in a way that undermined the narrative what is the narrative uh in the world of tooth and tail an unspecified calamity wipes out all the crops causes a massive food shortage and uh and all animals even herbivores they're forced to eat meat to survive but then they all kind of decide they like to eat meat they prefer it so this war is ultimately about who gets to be the people and who has to be the meat it's modeled loosely around conflicts the dawn of mechanized warfare so world war one but i think more specifically the russian revolution so this is not a happy place but the title approach that was being used when i joined the project kind of undermined all that we're saying it's about terrible social upheaval and this kind of unprecedented carnage and a genie that's been led out of the bottle but um what we're showing you is like marble madness special edition re-release if you're too young to know what that is that's a game that like caveman played going to dinosaur school uh just look it up it looks just like that uh okay so another problem there was no achievable end goal defined basically there is no like actionable mock-up the mock-ups that were made looked super nice but these were really paintings and not proper mock-ups given the tech that the game was built on there's really no way of recreating this in game so you're still left with this big open question of like okay well uh this is great but what does the game actually look like the arc lacked a solid direction to head in um another problem was that there were a few thoroughly tested arp pipelines in place now um pocket watch likes rolling their own engine and the engine had all the support for like the major features you would expect it had actors and it had objects and it had a bunch of environmental stuff and it had particle effects and all this kind of stuff but nothing was really documented and uh an experienced artist never really put all of that stuff through its paces so there were big questions about where the weak points in the tech were um and then finally the last little bit of critique there was no art director so andy schatz the um the again the main pocket watch guy he was playing art director while he was also being a designer and running a business and being a programmer and even in the indie space which is kind of known for people wearing multiple hats i think that was one hat too many i think another thing though was uh andy does have some experience as an artist he actually has an art degree and he he paints and all in a bunch of stuff he's actually pretty decent painter but he hasn't spent years solving visual problems and if you've got a stack of visual problems to solve it's going to be most efficient to have someone who does have experience doing that trying to work through that stack another thing i did was i took stock of um the things pocketwatch had going for them and the top of that list was the painter and the project jerome jacinto jerome is awesome the art all the art that i just spent 10 minutes critiquing was made by jerome but jerome is an incredible painter he can paint circles around me all day before i even get out of bed uh and he's just awesome to work with but at the time i came on board i think he was still pretty green in terms of being a game artist he knew how to make a still image that would blow you away but hadn't yet developed a ton of experience kind of wrestling with the complex um changing informational landscape that is a game but it was pretty clear that with just even a little bit of direction he was going to be like a super weapon and that that was the case so lessons i'm going to be calling out key lessons at regular intervals and and i'm hoping my aim is that these lessons are broader than the talk itself it's they're things that you can apply to all kinds of art making not just like 2d stuff but even hopefully other disciplines so here's the first batch even if you are crunched for time allowing yourself to stop and think kind of about what you're doing usually ends up paying dividends in the in the form of um in productivity you end up working faster after you have that time to think and you end up making less mistakes um i know it's super stressful potentially because stopping to think does not feel as much like work as like trying to just like brute force your way through an asset list but in my experience it has always been worth it another one fancy tools are awesome and we have tons and tons of fancy tools in 2017 but i won't say they're useless but they're not as useful as they could be if you don't understand visual design basics so if you want to kind of develop these skills uh skip the you know like how to model a soldier and blender tutorial and and go for like basic color theory or understanding anatomy or something like that and even if you aren't an artist checking out something like that you're going to be able to talk to artists way more effectively uh and maybe prototyping even goes better so i think it's a good exercise and then three it may seem like semantics at first blush but i think there is a really important distinction between being a game artist and being an animator or a 3d modeler the artwork in games is i think in its purest form uh visual communication of the gameplay yes it's also there to kind of like get people amped and like keep people there once they're there but you can kind of take those things away and the art still works as long as it's communicating the design there's another discipline that is primarily concerned with visual communication of information and that's just graphic design so besides those um art fundamentals that i was just talking about you know learn about branding and typography and like dive down into this kind of stuff do research about subway maps or signs in like tourist areas and then try and draw the connections between what makes that stuff successful and what makes an interesting character design or an interesting the design of a locale or something like that okay so we knew what the problems were uh and now it was time to start addressing them the very first thing i did was i made a new mock-up that addressed those formal issues i just talked about while also trying to reinforce the game's narrative i kept it super simple at the beginning because i wanted to be sure we could easily talk about everything that was there i wanted all of these early ideas to be properly critiqued i wanted nothing to be lost in the mix and i wanted everyone to understand like why we wanted to do each one of these things basically a metaphor that we can kind of think of i'm trying to like lay a really solid foundation here and maybe like sketch out a four floor plan but in no way am i trying to show a finished house i think there's a lesson in that baby steps i'm going to come back to baby steps again and again throughout this career it ensures an entire team especially a remote team i was remote pocket watch games is in san diego i live in maine it ensures that an entire team can kind of keep up and it stops problems like small problems from becoming compounded i think also if you ensure an entire team is moving together you're kind of sharing the experience of making that game and i think you're more likely to be on the same wavelength in terms of communication so this mock-up was made with actionability in mind it only included artwork or features that i knew we could add uh there wasn't any bit of art here that made me say like oh well this is cool but i don't know how we're gonna do this yet no new questions only answers uh let's go through some of these uh things bit by bit so one suggestion was instead of those like grid blocks then the checkerboard that the game was using before um i suggested we do a fill of like much larger textures at the base level of the world kind of underneath the cliffs and underneath the units and stuff and that was to help reduce patterning make the world feel a little bit more organic and get rid of that checkerboard effect i suggested we use a shader even just like an old-fashioned overlay image to shift the color of the train so as to get a little bit more color variation make it feel a little bit more organic and i suggested some simple atmospheric overlays so cloud shadows you know that scroll across the ground and animated sun rays on a post process tint this was to help push the value range and make the whole scene feel a little bit more dramatic there's also some changes here meant to solve uh some like mundane but critical problems so an example uh z fighting or like bad sorting was a problem with the way the the old terrain works sometimes units would render behind squares uh that they were in front of and vice versa so i suggested we paint terrain are we rather pink cliffs and ramps in such a way that you can never actually get behind them a lesson for junior art or really actually all artists that if you can reduce like bugs for coders they will love you forever and do that just do that so the mock-up was well received and i don't want to put words into anybody's mouth but i think everyone kind of felt reinvigorated a little bit and it was around this time that andy said like paraphrased um i i always imagine this as a pixel art game can we do pixel art is that even a good idea would that work how does that change the schedule can we do it i loved that suggestion because command and conquer won the very first one was the game i was obsessively playing when i realized it was like someone's job to make games and like holy i could be one of those guys uh to make a strategy game with big chunky pixels was like a pilgrimage for me uh but more important than that um personal excitement uh was the fact that moving tooth and tail to a pixel style kind of spoke to the game's soul better after all it's a bit of a love letter to those classic 90s strategy games while also being a little bit of a love letter to some classic arcade games and like arcade culture it also simplified animation production tremendously before it was a bit of an open question of how we would actually produce all of the animation in the game uh it would probably require a new tech or be expensive would we use like a 2d rig system would we model and animate everything in 3d would we do like high-res uh hand-painted uh animation nope we would use good old-fashioned hand-drawn pixely sprites uh we could focus entirely on like broad gesture without getting caught in expensive details relatively nice and easy you know compared to those other options and it works with our existing tech great so the first pixel mock-up was a hit and again while it was a little bit thin at the start it proved to be like the perfect foundation on which to build but i will get to that in just a second because let's do some lessons first if you pick a an art style that plays to your kind of your team's strengths and speaks to your core experience you reduce the number of visual problems that you actually have to solve it makes you in your team saner because you're not trying to figure out a way to like say x with a visual language that just isn't well equipped to say x in our case pixels spoke to the soul of the game really nicely it solved a bunch of technical problems and it was fast to produce so it was a plus for us and uh two um a mock-up is a tool and i think in some cases in my experience maybe even most cases it's the most valuable tool you have it's not just a painting or a fancy render it can often give you a leg up on production too because like the assets you make in the maca can just be moved into a build so in my experience it's always been good to spend real time on it and not think of it as like oh we just have to get this done and then we can start making the real art okay let's keep going so we had a mock-up we had total team buy-in we were all excited about pixels there we go uh at this point we started pushing forward on multiple fronts so jerome and i dove into character design we did just iteration after iteration of of line art this is like the tiniest tip of the iceberg i would need all of the walls to show everything we did it was exhaustive uh jerome himself did a bunch of like paintings exploring a bunch of different fun pot none of these are fun but different but fun potential scenarios and emotional themes i want to make a game just about this dog getting his hat shot that doesn't show up in the game but it should probably so the end result of all this work was a nice set of guidelines um that dictated the visual design of our units so units in a given tier we've got these three tiers units in a given tier um all stand in different poses or have drastically different equipment or weapons so as to help the silhouette be as informative as possible reptiles are primitive they have less clothing and they generally have melee attacks mammals wear more clothing and are generally ranged scale clearly denotes combat effectiveness one thing that i've learned is that my job is a visual designer it's not just to design stuff that like looks cool it's to design stuff that allows for a suspension of disbelief and creating a consistent visual language something that has order and patterns and stuff really helps do that the mock-up itself became something of a living document so if anyone had any idea for art it was iterated a half dozen times in the mock-up before we ever actually put anything in the build one of the best pieces of advice i ever got from my first art director was it's cheaper to iterate on paper and i think this is a version of that it's maybe also worth mentioning that all of this work i've described so far happened pretty quickly about the first two months one thing that helped us a great deal as we were doing all of this was a slightly exhaustive visual communication among the team little story boards if you will that kind of describe tech needs or how a sequence of events should work since we didn't have a lot of time we couldn't afford to lose time due to miscommunication so if we're ever kind of talking about a complex topic we would do something up like this to make sure that all of these ideas were like completely thoroughly understood i came to understand the value of this though through the context of my wider career i've worked with like a lot of different kinds of people and a lot of different spaces um certainly teams like pocket watch but also teams in academia research teams marketing teams and i've learned that kind of everybody communicates a little bit differently so working to convey a single idea in like multiple different ways helps ensure that that idea is going to be understood regardless of what kind of communicator is on the other end taking a little bit of time to communicate clearly in my experience helps prevent miscommunication which can be like super costly sometimes but uh we eventually had to move from these like nice looking mock-ups and start putting the stuff in the game and wrestle with like the two or three weeks where the game looked like an awkward teenager which is what this is right here uh so at this point um i proceeded to create at least one of every major um asset type in the game so at least one commander at least one unit at least one building um and uh the goal was twofold uh one to get like a firsthand um a little bit of firsthand experience and understanding of like the assumed art pipelines um and so we could identify the weak points in that tech but also i needed to know how long it was going to take to make each one of th each one of these things because um we didn't we needed to build a schedule that was based on something besides just uh like uh gut estimates and i there's a lesson in that um uncertainty is in my experience just about your worst enemy uh experience is always better than an educated guess and the longer i have let blind spots live the more time they've cost me in the end so i again in my experience it's best to remove those blind spots as early as possible so at the onset of these tests i was thinking like okay great this is a pixel art game i can do everything in photoshop i can have a nice neat like single piece of software tool chain and since we're dealing with tiny resolutions i can basically just make like a super small psd like 480 by 270 pixels and just paint all the environmental assets right in there without worrying about blowing through memory uh everything would look consistent because uh it's all like just painted right there in the same scene and in the end that is how we produce all the environmental assets in the game i have a psd with no joke between two and three thousand layers that just has bam every environmental asset uh that's in the game but these early pipeline tests showed us that um animation was still going to kill us even even after the move to pixels it was still going to kill us now it's worth mentioning that one of my goals for the animation um was that i wanted really lively kind of organic nuanced animation and not like the limited frames no in-between approach usually associated with traditional pixel art and the reason i wanted to do that was because one of the conditions of moving to pixels was that we had to be sure we did it in a way where we couldn't be written off as just another indie pixel game i hate this animation so in theory the move to pixels actually made that goal of like really nice animation more achievable like i said before we could focus entirely on big broad gesture we didn't have to get caught in details um we could have all the in-betweens i i wanted we could do it right at chickadee we didn't have to like hire an army of french animators but animating in photoshop for me is like so cumbersome and i know it is a bad artist that blames his tools uh but there's no way to like quickly move between frames and like and i'm talking about like like the rough animation not not painting frames if it just feels like everything takes more than one click it's really hard to nail that initial gesture and that makes the final results suffer so i need to figure out a new tool chain at this point and realistically i needed to find a way to cut scope too because there was just so much animation in the game so just like before i made time to like think about the problem before me i allowed myself to consider ideas um that seemed bad this isn't actually me thinking i accidentally deleted like four hours of work and then i just laid on the floor for five minutes so around this time though while i was trying to solve these problems a an insane opportunity insane fell into pocket watch games collective laps hold on water okay so um i got the call while i was at lunch with my wife at the bar underneath our studio and andy asked this is crazy but do you want to be in a coors light commercial so evidently coors was producing a new marketing campaign based on like makers and artists and musicians and i guess how the cool crisp taste of the rockies inspired them to make cool stuff i swear this is true uh the the premise of this commercial was that there's this team of like young dudes in a bar and they're like scrolling on napkins and getting all excited and then it shows them like i guess in their office and there's like big character drawings and stuff and they're like oh wow i'm making the game and then uh there's like a launch party in a bar and like they're showing the game up on a tv and it's like whoa we made the game and just everybody's drinking coors light because that is that's like a very realistic thing that happens to all of us uh but here's the kicker uh we only had a fraction of the art in the game we still had awkward teenager art going on um so i had to do like a little mini crunch to make like a fake gameplay loop in a fake title sequence so uh here's the here's the title sequence it looks pretty good right uh and then here's the gameplay lube so um this is like squirrel d-day basically and it never gets better uh so uh so the commercial was actually shot and then coors decided to change direction the direction of the whole campaign and the whole thing was scrapped um and i because you need closure uh i'll tell you that i declined to be in the commercial because and this may be shocking to all of you i don't drink a lot of light beer and it felt really dishonest to like you know promote it uh but the gameplay loop was like the most perfect albeit like super random task to be required at that moment it was a great proof of concept for a few ideas that i was kicking around in my own head and they turned out to be game changers other game changers for the rest of production so the big one was the unit animation pipeline uh given that i only had a week to get this done i went with a pipeline that i knew worked it was a little long and a little clumsy but i knew it worked what i would do is i do the rough animation and flash i'd export it out with without anti-aliasing um i'd rebuild the sequence in photoshop and then i would just paint the frames uh this was the same pipeline i developed to create a trailer of a little unreleased um game that chickadee was making in-house it was clumsy for that uh because the characters were bigger the sequences were longer the shading was a little bit more complex and had to be tracked basically there's just like more of everything if i got to photoshop and i realized like i wanted a core change i would kind of try and make it happen right there but usually i'd have to go back to flash and that means any painting i'd already done would be scrapped and it was this like really cumbersome back and forth thing that just took a lot of time but for the characters of tooth and tail like the squirrel comes in at like 10 by 10 pixels it's tiny this whole thing was lightning fast for this stuff there wasn't really a need to ever go back and forth as long as i i really tried to capture like the dust of the animation in flash any changes i needed to make i could just make in photoshop so on the surface it seemed like this was going to be clumsy but it turned out to be pretty efficient and then once we took it to production i added a few steps that again it made it seem like it was going to be slower but it turned out to be even more efficient so in production i would create the rough lines and flash and then i would test those in game and then i would clean them up a little bit in photoshop i'd do some like simple color fills and i would test those in game and then i would bring them back i'd do my final shading passes and i'd be left with my final artwork so i mentioned before about the benefit of um baby steps that they stop like small problems from becoming calm patent and turn into much bigger problems richard williams who is like a master animator he wrote this book animator survival kit and he talks about the long shortcut he says the long way turns out to be shorter because something usually goes wrong with some clever rabbit's idea for a shortcut and it turns out to take even longer trying to fix everything when it goes wrong this is some of my absolute favorite advice i love sharing it in my own creative life i've taken it to me and important stuff usually takes time to do so don't make it go even longer by half-assing it so in the case of the animation pipeline baby steps maybe took more time up front but i no doubt save tons of time that like reshoots would have taken had i not been so cautious we had another win for the unit animation in terms of a scope cut so originally andy envisioned four directions of movement for the units so diagonally up down left and right i suggested keeping them in profile would help the silhouette read more clearly and consistently while also reducing the amount of animation we had to do but there was a little bit of a concern that maybe that would just look wonky like a uh unit staying in perfect profile walking around a you know a more 3d world but the commer the commercial proved that this approach was going to work just fine um additionally i myself discovered that i really didn't need as many in-betweens as i initially anticipated um so you know again test your assumptions i was wrong i'm happy i did my test another big gain from the commercial was the discovery that we could use 3d models to create pretty decent pixel work now when i first thought of this idea i laughed at how bad it was because it is a bad idea certainly we could use 3d uh models and render sprites but trying to use 3d uh models to render sprites that looked like they were hand painted seemed like a really roundabout way of getting a result right just paint the sprite but just like with the character animation there's a little bit of experience from my past that i could draw on so one evening many years ago i was fooling around and i made these little 3d shapes and i rendered them at a super low resolution and i thought like oh that's kind of funny they look like pixel art but i made them in macs and i tossed them on my tumblr and that was that but because i had this tiny bit of experience in my back pocket i thought okay well this is probably going to be a waste of time but the fact is i need to figure out a way to render or to get like 60 or 70 frames of like slow windmill rotation animation and i need to get it quickly so a proof of concept will probably only take me an hour and then once i fail at that i can kind of just get it out of my head and figure out like the right way to do it but in an hour not only did i have my proof of concept it was just done that's all it took which was awesome so i'll walk walk you through the setup real quickly just in case anyone's interested so first i made a camera that was angled to match the in-game camera then i tossed the windmill tower which wasn't animated that's just on a plane um in the background and the plane is aligned to the camera i created two versions of the windmill blades one is aligned to the camera which kind of gives it a weird angle it makes it hard to work with but then i had an instance proxy off screen or off camera and that was a line to the world which was a lot easier to work with i unwrapped it i gave it a tiny tiny resolution texture i turned off filtering on all the maps to maintain those chunky pixels and then when it was time to render i just turned off anti-aliasing um and set set it to like a really tiny resolution and that was that i got my 70 frames of animation and it only took an hour which was awesome again the process actually worked so good that we just kept it around for the um the final windmill this is the windmill that's actually in the game um i'm gonna jump ahead a little bit um but this is all related this workflow found an even more important use down the road which is the for the commander flag now there are four commanders each one has a different flag and each commander has about 80 frames of animation and that flag is moving around in all different ways uh for each of those frames there's no duplicates so in the original commander test which is what this is i hand painted those flags and it took forever in fact how we were going to get this done was the only open question we had left going into production i just kind of assumed i'm gonna have to hire someone to do this part of it because i can't be tied up for like four weeks just painting fabric but i was also a little bit concerned about the fragility of it right like if we went ahead and did all that work and then we realized we needed it like a core change to the animation that is like a ton of time just tossed out the window and again we're working on like a really tight schedule uh but the 3d pipeline saved us here and so again at first blush i thought this probably won't work the commanders are way more complex than the windmill they're moving around in all different ways but i decided again it's worth a proof of concept i can probably get something going maybe not even something that looks good but i could figure out in about an hour or two is this worth pursuing um and after about an hour i discovered that yes despite that extra complexity we could get a 3d flag to track with the original 2d flag at this point all the commanders had just like a flat shaded flag in place that was that was properly animated and you can see here it's like super simple um it's like just kind of a low density mesh the bones don't even have controls i'm sorry animators i know that's super super bad and after a couple more hours i figured out a material setup that yielded pretty good results so i had two uv sets on the flag the first one held a static texture the chunky pixels kind of just like the windmill and the second one had um a bump map on it that was animated to scroll uh from the flagpole out to the end of the fabric um and that was used to kind of create these like tiny undulations that would have been time consuming to do like in a rig you combine the physical animation of the flag with that texture animation and you get the final composite which uh you know look that works pretty good um here it is in game now um after a couple more hours of of lighting and tweaking values here and there kind of trying to fine tune it this is what we got so um because we're all kind of used to looking at um a lot of game art and working with this kind of stuff um i'm sure you can all look at this and say i mean that's pretty cool but it doesn't like perfectly match the hand-painted artwork and that is true but uh i am a pragmatist and i believe in the admittedly unromantic but i think still um powerful and human and kind of poetic notion of good enough good enough has a bad reputation i think uh when i say good enough i think maybe a lot of you can imagine a time when when you've had to say like okay well this is we have to ship so whatever good enough but i think good enough is like more measured and powerful than that so as i said before uh i'm a college educator and one of the things i tell my students is not to stress out too badly about finding a job in order to get a job you really only need two things you need to have the base level of skill and just don't be an you're going to enter the industry as a junior artist and people know when they hire a junior artist there's going to be some education that needs to happen right you're not expected to do the work of a lead artist right out of the gate maybe as a student you don't even realize there's a difference between those two things you know but don't worry because if you try hard if if you put the hours in learning those core skills you try and develop your ability to solve visual problems you get out and do some networking and you try and be good person don't worry you're going to be good enough so what's the real philosophy of good enough then good enough is about looking at the whole package and weighing all options i tell my students i've never hired the best person who applied for a job when i've been in a studio and in charge of hiring artists i've never hired the best artist but i've always hired the best fit certainly how well someone can practice their craft that is a huge part of it i i don't want to mislead anyone but also what kind of person are they are they nice to be around do they express their ideas clearly are they a good problem solver solver are they self-reliant are they funny are they nurturing are they responsible or are they just an so back to the flag uh this process turns uh five weeks of painting by hand into like just a week of work to get all of them done we can do it all at chickadee and spending instead of spending money on an outside contractor we can if we need to do a retake we can build off of what's already there if we need to add more animations in the future again we have our base so the slight mismatch in visual style uh is totally fine when weight against those gains and what's more the 3d flags actually made like an awesome feature possible that would have been totally impossible otherwise using this approach gave us a really easy way to do decals custom decals on the flag because we have a 3d model and a camera we know where things are sitting in space so i could render out a sprite sheet of a flag that was basically just the screen space normal map andy was able to use the sprite sheet to dynamically place any random image on the flag in real time here we have a little monaco guy so players now have like a way to make like custom banners for themselves or their clans i don't know if clans are still a thing they were when i was younger it's just going to be used for memes probably but but still it's pretty cool um and just think the the decals on the flag the whole 3d pipeline that the um the 2d animation pipeline all of these beautiful trees grew from seedlings nourished by the crisp cool taste of the rockies lights my talk should be sponsored by coors and it's not okay but let's do some lessons so i won't say that bad ideas don't exist but i do think that most ideas probably aren't bad uh they're just ideas they're maybe not widely considered and therefore maybe they're a little scary but i think that also means that they're not obvious and therefore they're loaded with potential um so when looking at like an unconventional idea resist the urge to say we've we're under too tight a deadline to consider that crazy stupid idea if you can do a proof of concept in a short amount of time i i think you're crazy not to do that and make time for random experiments outside of your main work i am admittedly terrible at this i have an iron work ethic and a crippling sense of responsibility i'm really fun to be around uh so i have a hard time to just allowing myself to just sit and make art um but if i hadn't been like just fooling around in max all those years making stupid little primitives render as pixel art while my wife watched like i don't know like um pretty little liars or something else that's like great i probably would never have considered the 3d for the windmill and the ramifications of that would have been like astronomical so on to the last leg uh we it was early summer of uh 2015. we had our pipelines we were on schedule we pushed ahead uh we announced the art style switch along with a name switch in the august of 2015 the game used to be called lead to fire now it's called tooth and tail here was a game that was a visual homage to the greats of the past but i i think was also a bit of a fresh take and had some new ideas in terms of gameplay i suppose it's ultimately a subjective thing uh but i do think we made a pretty good case for why we weren't just like another indie pixel game the clouds and and sun rays and uh and water were all uh dynamic and shader driven the water actually was like basically a 3d water shader crammed down into 2d and the waves were like driven by dynamic wind forces and props to andy from making all that stuff happen because i originally said oh just make it an art and he was like no it has to be real and it looks dazzier it looks easier than if we tried to bake it in art um so as i said before the game was always about a blast to play but the kind of inconsistency of the old art um got in the way of that uh or got in the way of people discovering that rather so with that barrier removed the press was more interested and more people started joining our alpha community this is my favorite slide uh in the fall of 2015 chickadee picked up a couple interns that were like instrumental in helping us hit our art target harrison brown and samar makuk so together along with jerome we all pushed ahead on on the art so let's just take a look at some shots in no particular order we iterated the environment artwork so we're using color and value here to indicate collidables and elevation we animated all the units this is just what i could fit into a single slide there's about twice as many units are points of high contrast in vivid colors so they pop off that terrain we had time for some winding iteration on a few units i've learned that it's always good to plan dead space into any schedule because not everything is going to go perfectly you're going to need a place for that those mistakes to kind of fill into we animated all the buildings this windmill right in the center is like my favorite thing from the game uh i made a point uh to pocket watch well it's worth saying that all this animation was done pretty late in production i made a point to pocket watch that we should live with the static versions of the units and the buildings as long as possible so that the inevitable like ebb and flow of the design wouldn't relegate a bunch of um animation to to the trash uh we wanted to wait until the design was more or less settled before like diving into this stuff again because we had such a tight deadline jerome painted amazing portraits for each uh unit and each commander also did a bunch of marketing shots that aren't shown here we lost a chunk of time to ui iteration and that's kind of a pretty common thing ui always takes a long time but i i don't think we did ourselves any favors because we never really had a solid plan or took time to think long term as we were doing this stuff it was kind of like oh okay well once we make this change then it's going to be perfect and then compounding that poor mentality was the fact that our tools were a little cumbersome didn't really uh they didn't really allow for like um rapid iteration so uh you know because we were worried about taking too much time we really didn't do the proper legwork um and it ended up taking away more time in the end we we should have done the um the long shortcut right and we didn't and and i think it cost us but hey look at this particle effects that was kind of like a downer slide and now we have like a really zazzy slide right here uh so there's there's tons of it's a it's a game about battle there's tons and tons of effects and they were a lot of fun to do uh so art slept a little bit we didn't hit the end of the year but we did hit like the second or third week in january and then we shipped shortly there after just kidding we worked on the game another year and we did that because game development uh so before i go any further i should be like absolutely clear from this point on this is like personal reflection um this isn't pocket watch it's not even really chickety it's just you know one guy who makes games kind of reflecting on that stuff it's a little bit like a like a personal post-mortem except the game's not out yet so it's like a pre-post-mortem i guess i don't know uh so the extended development was seeded i think by our improved standing with gamers and press uh when i started on on the project i think there was a little bit of a fear among the team that this was maybe a niche game and it might bomb the increased visibility and excitement reduced those concerns but replaced them with the worry that maybe a minimum viable product wasn't going to be enough and then this guy joined the team and he began wagging his jaw uh some of that wagon was pretty decent it made decent sense right um it felt logical some of it was a little more superfluous if you've been in games if you've made a game even as a student you kind of know about the feature creep a lot maybe most of his ideas are cool but he talks and box bullet points he doesn't really consider objective merit or time but he appeals to emotion like a great salesman and it's really hard to not listen let's take a look at the weather system as an example the weather system allows us to do surprise different kinds of weather but also different times of day and it makes the screenshots look super zazzy they look really really nice but it dismantles the visual hierarchy a bit it it flattens values and homogenizes color it actually makes the game harder to play as a visual purist i lament that loss of clarity but at the same time it makes for a more marketable product as far as screenshots and video are concerned it looks fancier so was it worth it i don't know i suppose it depends on what metric you're measuring but our process i think was undeniably messy we took an idea suggested by feature creep we didn't do our homework beforehand to figure out the true scope of it and we implemented it before we had a mock-up we basically tossed aside a bunch of the very practices that led to our previous success this wasn't a brick laid in surface to a carefully considered long-term plan or even really a poorly considered long-term plan it was just a brick that we kind of liked the shape of so who knows if we actually needed it to finish our house it may be made for a better product but did it make for a better game i don't want to give the wrong impression either we were all working really hard during this time we were adding new features we were showing the game off at conventions uh we were considering publishing and platform possibilities seasons passed um i think we're all just you know like trying to make the game better bit by bit and that's cool that's a very organic way to approach game development and i think there are times when that's exactly how you want to approach your craft but in comparing the first year i was on tooth and tail to the second i think all evidence points to the fact that we did our best work when we were under a deadline the initial you know get this done by the end of the year deadline was self-imposed but we stuck to it and helped focus our decision-making we didn't have time to be wishy-washy the week-long crunch for the course commercial forced me to quickly test a few ideas that turned out to be game changers now obviously the reality of game development is that it's time consuming um i get it and i'm not trying to argue that like through the like magic of project management you know you can find a secret shortcut uh from the kitchen to the study that was about clue a little esoteric i know but i really wanted to put it in there but i will say that smart consistent project management in this case you know slowing down to think about our approach having a clear goal in mind uh tossing dead time into a schedule for when things don't work out basically all the stuff that i just spent 45 minutes talking about it allowed us to get about 12 months of artwork done in about 10. we were like art machines that first year in the second year we were way less efficient we wasted time now i'm sure many of you can think of inefficient development cycles and think of like how expensive it is like oh that's extra monthly burn and that's true but i think that's a very business oriented way to view the downside of poor planning i think the steeper cost is far more human the monetary costs of your development are recupable but your time is not so something to think about if through thoughtful planning and disciplined management you were able to release a game three months earlier than you would have otherwise that's effectively free time what would you do with that time would you um would you prototype a new game would you dive back into it would you learn a new skill or gain an experience that would inform your future game making or maybe you would just go outside once in a while and spend time with loved ones again i get it development is just hard and it takes a lot of time but poor planning is just about like the dumbest uh most pointless most avoidable just shittiest excuse uh for wasting time um you're not getting it back none of us are uh the time we spend on our games comes from some place it comes from every other experience we might gain if we weren't doing that simply put to borrow a phrase from andy nguyen one of the other pocket watch folks he talks about maximizing value the man likes to find a good deal i think simply put it's tragic to not maximize the value of your life but i'll tell you what this is heavy right we thought we were like learning about pipelines and stuff and it's like what's this guy talking about now uh i don't want to come off as a bummer or defeat us or anything like that because tooth and tail actually in my eyes is already an incredible success and we haven't launched yet but it's it's a success not for any of the reasons i've talked about in the last hour it's successful because it brought people it brought other human beings uh it brought them together um even before launch tooth and tail um has had us an astounding community and this is largely thanks to the uh cultivation uh efforts of of andy winn the um guy that i just mentioned um this is a community that's hundreds and hundreds of people um before we ever launch active players and i don't say that as lego it's a stats thing it's it's um this is these are people all over the world sharing their excitement and happiness and they're learning together old players help new players learn the game they invent new strategies and like record them and put them online and it's like oh my god do you see what that guy did there are people who modded the game before it's ever been released there are people who who are learning how to make games by modding our little unreleased game that is mind-blowing to me one community member captain fog that's him right on the second third of the image right there uh he came out to help us at pax east he was a local there in massachusetts he did like all the grunt work um like grabbing waters and like grabbing sandwiches and stuff like that and he demoed like a pro he was awesome he was my mvp of the week uh so we made a thing that was a force of good in the world and that is my personal best case scenario for what defines a successful game wait against that i feel pretty comfortable saying that all of our development practices warts and all were probably good enough that's my talk thank you um you have any all right well uh thanks for attending everybody fill out the thing because it helps this whole thing go better oh wait do we have a question thank you yeah just had a quick question about the flag stuff yeah sure um just how dyna dynamic did you have to make the flags as far as like how they moved around with the characters are they like separate sprite sheets that are oriented around the characters or so so that's a good question so the question is are the flags different sprite sheets um uh from the from the rats no uh the rats and the flags are all the same sprite sheet although when i rendered uh the flags the rats weren't included in that render so i basically did a pass of the flags and then i had the original rat animation with the flagpole and stuff and i just smashed them together in photoshop and saved out the final the final image the final spreadsheet yeah thanks yeah hi my name is chris and i'm from indonesia actually uh i saw your game on pax east and it's amazing oh thank you and um my question is you seem when when you change the art style it seems like there there's some changes in the code as well like in the grid system and everything yeah and because i think i see that you have some technical knowledge but what if there's a for example like an indie team and the artist basically wants to change the art style but he doesn't know the technical side of it how do they communicate with the you know programmer and how do they convince that that's a really good question and that's maybe a big question um so um the question is uh how do you make these big broad changes if the artists themselves don't possess that that broad technical knowledge uh so in tooth and tail um andy schatz is an amazing um graphics programmer that was one of the team's strengths so i knew that i could make requests that were moderate and he would be able to get them done in a reasonable amount of time if you don't have that or well i don't possess those skills myself right i can't code up shaders that's something i can't do but in all of these uh requests that i made i did those kind of like um storyboard uh images of like okay here is exactly what i want and like uh a lot of those images were just here's how i made it in photoshop step by step like oh i blocked out like this black portion and then i blurred it and then i did like a weird distort thing and then we use that as a mask and and fill in a different texture and that's enough that was enough for andy to say like okay cool i see what you're doing here and i know how to translate that to code um i think really i don't know if it would work for every team but the way we did it was just being really exhaustive with communication and kind of basically drawing little pictures to aid that i'm not sure how it would work with other teams though it really i think depends on the dynamics of your own team all right thank you everybody

Channel: GDC

Views: 60,769

Rating: 4.9579477 out of 5

Keywords: gdc, talk, panel, game, games, gaming, development, hd, design, monaco, tooth and tail, steam, art direction

Id: rSk8adHotDA

Length: 64min 52sec (3892 seconds)

Published: Fri Dec 04 2020