Building Customizable Characters for Destiny

Summary

The presentation by Scott Shepard, Bungie's technical art lead, delves into the intricate process of developing customizable characters for Destiny. With insights spanning from conceptual ideation to intricate design mechanics, Shepard outlines how Bungie's approach has evolved from their previous work on Halo. The focus on player-driven storytelling in Destiny necessitated a deep customization system allowing players to express individuality. Shepard expounded on the game's character archetypes, the technical challenges of designing modular components, and the iterative process of refining tools to facilitate more effortless content creation.

Highlights

• Scott Shepard, Bungie's technical art lead, led the presentation. 🎮
• Destiny allows for deep player customization. 🛠️
• Three character classes: Hunters, Warlocks, and Titans. 👾
• A modular system helps in designing diverse character gear. 🧩
• Challenges in maintaining creativity with modular designs. 🎨

Key Takeaways

• Destiny's characters are designed with a focus on customizable storytelling. 📚
• The modular system allows for diverse gear designs while maintaining artistic control. 🎭
• Tools were developed to optimize creation and maintenance of character assets. 🔧
• Design iterations were crucial in achieving a flexible character customization system. 🎯
• Flexibility in design allows Bungie to expand content over a decade. ⌛

Overview

Scott Shepard, the technical art lead at Bungie, dives deep into the rich process of crafting customizable characters for the game Destiny. His talk at the Game Developers Conference highlights the journey from developing the iconic Master Chief in Halo to creating a new world where players can tell their own stories. With Destiny, Bungie aimed to craft an experience where each player could stand out through personalized characters, a significant shift from the singular hero model of earlier games.

The presentation uncovers how Bungie's design philosophy emphasizes the player's unique experience by providing a highly customizable character system. Three primary classes—Hunters, Warlocks, and Titans—each offer distinct looks and functionalities. The development included modular systems, allowing artists to interchange gear parts fluidly, finally enabling a vast array of possible character designs while maintaining a signature Destiny aesthetic.

Shepard also addresses the challenges and solutions encountered along the way. Bungie's iterative design process, the introduction of tools like Gear Manager and Mantini, and the innovative die system showcase the technical expertise and creative flexibility behind Destiny's character customization. These insights reveal Bungie’s innovative steps towards creating a sustainable and expandable gaming universe.

Chapters

• 00:00 - 01:30: Introduction and Background The chapter provides an introduction and sets the background for the presentation. The speaker begins with casual remarks, checking audio levels and reminding the audience to silence their phones, humorously warning of mockery if interrupted by a ringtone. The chapter also mentions an electronic survey to be filled out by the audience at the end of the presentation, emphasizing the importance of feedback.
• 01:30 - 05:00: Destiny's Character Design Philosophy The speaker introduces himself as Scott Shepard, the technical art lead for the 3D department at Bungie, and mentions his long career in the games industry since 1999. He expresses his willingness to receive feedback on his presentation and humorously notes his tendency to speak quickly when nervous, assuring the audience that there will be time for questions at the end.
• 05:00 - 09:00: Character Customization Goals and Archetypes The chapter explores character customization in video games, focusing on Bungie's work, particularly on Destiny and Master Chief from Halo. It reflects on the influence of past projects on future endeavors and character development across different game genres.
• 09:00 - 18:00: Technical Approach to Character Customization This chapter delves into the extensive history and cultural impact of character customization in the gaming world, using the example of a well-known character with green armor and an orange visor. This character has permeated various media - from games and books to films and merchandise – creating a lasting impression. The faceless design of the character was a strategic decision to allow players to project themselves onto the character, becoming the hero of the saga. This approach is highlighted as a significant touchpoint in character customization. The chapter transitions to question the future direction of character development while maintaining this impactful player connection.
• 18:00 - 27:00: Tools and System Implementation The chapter titled 'Tools and System Implementation' explores the unique approach taken in the development of the game Destiny. Unlike games that focus on a singular hero or storyline, Destiny is designed to create an immersive world where the player is the hero. The development team aimed to learn from previous successes, such as the creation of the Master Chief in Halo, but they sought to develop a distinctive experience for Destiny. The guiding principle throughout the project was to build a world where players could craft their own stories and continually return to for unique experiences. Ultimately, the hero in Destiny is each individual player, emphasizing the personalized and customizable nature of the game.
• 27:00 - 35:00: Challenges and Solutions in Development The chapter discusses the importance of player agency in game design. Developers focused on creating a system where players could feel like they are creating their own character roles, rather than being given predefined roles. This required enabling players to discover and shape their unique avatars within the game's world, particularly in the world of destiny. The goal was to ensure that each player could stand out and feel invested in the game, enhancing their gaming experience.
• 35:00 - 45:00: Color and Material Customization The chapter discusses the emphasis on providing a comprehensive and immersive customization experience for characters in a game, which surpasses previous efforts and industry standards. The focus is on allowing characters to evolve and grow, enhancing player engagement through dynamic customization.
• 45:00 - 54:00: Future Improvements and Conclusion The chapter titled 'Future Improvements and Conclusion' explores the broadening hero archetypes in the game world of Destiny, incorporating diverse characters like Space Marines, Pirates, Wizards, Knights, and Ninjas. The developers aimed to channel these power fantasies into recognizable archetypes that fit within the Destiny universe. This effort resulted in the creation of three main character classes: the Hunters, the Warlocks, and the Titans, which reflect the various stories players might wish to explore.
• 54:00 - 90:00: Audience Q&A The chapter 'Audience Q&A' delves into the design philosophy behind the character customization in the game 'Destiny'. It explains how characters are divided into four main slots: head, chest, arms, and legs, and how players are rewarded with different gear for each slot as they progress through the game. This gear not only alters their appearance but also their abilities, allowing players to customize both their visual and gameplay experience. The design maintains a balance between gameplay progression, visual customization, and aesthetic harmony. The chapter also mentions the use of basic silhouette shapes for each class, enhancing this balance and aiding in distinguishing the classes visually.

Building Customizable Characters for Destiny Transcription

• 00:00 - 00:30 [Music] hello can you hear me all right too loud too quiet everyone who was out last night is like too loud too loud um so a couple things before we get started uh please turn off your cell phones if it does ring I will mock you incessantly throughout this presentation um and at the end you're going to get that electronic survey we've all gotten them please fill it out uh let me know
• 00:30 - 01:00 what I can do better um if we need more content less content some sort of synchronized dance off whatever um and it helps me come back next year also I can talk a little fast when I get nervous so if we get to the end of this thing and you guys are like what the hell did he just say um we'll have time for questions so don't worry too much about it all right my name is Scott Shepard I'm the technical art lead for the 3D department at Bungie I've been in the games industry since about 1999
• 01:00 - 01:30 and have worked on everything from fantasy MMOs to survival horror the Sci-Fi firstperson Shooters of Halo but for about the past 4 years I've been focused on building customizable characters for Destiny to talk about Bungie's future we first have to talk about our past the master chief is one of the most recognizable characters of the modern video game era His Image was developed across 10 years of projects and
• 01:30 - 02:00 thousands of man hours of work he has appeared in Games books animated and liveaction films as action figures even on soda cans his green armor and orange visor ingrained in the minds of Gamers both young and old his faceless design allowed the player to really imagine themselves as the hero of the Halo Saga I call that out because that point stuck with us in particular so where do we go from here how do we on one hand build the
• 02:00 - 02:30 successes we learned from developing the master chief but at the same time create something entirely unique to Destiny who is Destiny's main hero well the answer is there isn't one destiny isn't about a singular hero or story our Mantra throughout the development of this project has been to create a world that people want to be in to go back to this is a place for players to tell their own stories in short our hero is you this
• 02:30 - 03:00 story is yours the idea of telling these player stories were the driving points for all of our decisions regarding our player characters it was important for us to feel important for us to let the player feel like they were part of this world invested in it that we weren't handing them a role to play they were creating their own role to accomplish that we needed a system that enabled the player to discover and shape their Avatar in the world of destiny that meant allowing each player stand out from others and encouraging
• 03:00 - 03:30 them to grow and change their character over time to succeed at this we needed a deep and satisfying customization experience that went well beyond anything we've ever attempted before and something we hadn't really seen in a shooter to date early on we talked a lot about power fantasies these are the heroes and adventurers that Inspire us at an early age this may be as simple as a game of cowboys and Indians as we are exposed to
• 03:30 - 04:00 different types of heroes it grows to include the likes of Space Marines Pirates Wizards Knights ninjas what have you we took these power fantasies and tried to wrap them into archetypes that fit well within our world what we emerged with were three classes we felt best represented the types of stories our players would want to tell in Destiny these are the hunters the warlocks and the Titans to begin our Explorations we
• 04:00 - 04:30 broke our character bodies into four slots the head chest arms and legs as players progress through the world of Destiny they would be rewarded with different pieces of gear for each slot this gear would alter their appearance and abilities allowing them to make changes choices that affected both their visual as well as gameplay experiences these four slots offered a balance between gameplay progression visual customization and aesthetic Harmony we chose basic silhouette shapes shes for each of the classes to help
• 04:30 - 05:00 make them more readable these shapes formed many of the ideas behind the initial Concepts and served as guides for ensuring each class maintained a unique identity to further separate the classes we added a fifth class specific gear slot the Hunter's cloak the Titans badge and the warlock's bond or armband the hunters are are Frontier
• 05:00 - 05:30 Scouts Scavenging and exploring in the world outside the city to bring back valuable technology the citizens rely upon they are hacker Cowboys bounty hunters and Recon snipers they carry themselves with style and Swagger Hunters scavenge much of their gear so we decided on an almost mix and match halfhazard look to their designs like the wandering Adventures of Old they wear heavy cloaks and combine utilitarian accessories such as packs
• 05:30 - 06:00 and straps with trophies of their Adventures both functional and fashionable we drew inspiration from survival gear like gas masks and hazmat suits so we chose to design their helmets with a focus on goggles and breathers their armor is lightweight and fitted to allow for mobility and speed is often comprised of a hard chest plate over a flexible Kevlar likee midsection they wear loose fitting pants with pockets and sturdy walking boots
• 06:00 - 06:30 the warlocks are our Warrior Scholars searching searching for answers to the mysteries of the traveler and a Mastery of its power they are Mystics Sears and wandering monks they are mysterious and secretive with the ability to harness the power of the traveler to devastating effect they wear long overcoats and robes and while they generally rely on their abilities more than heavy armor of other classes they still pack a fair
• 06:30 - 07:00 amount of hardware on their persons they're scholarly but also practical and they are no strangers to the heavy Firepower required to survive in the wild the warlocks had many influences and Inspirations from Wizards and shamans to the trench soldiers of World War I we tried to combine all of these elements while always maintaining an air of mystery about them and finally the Titans are equivalent to the Knights of old
• 07:00 - 07:30 heavily armored and bearing the emblems and heraldry of the city they protect they are noble and brave with a culture steeped in Tradition the Titans considered a great honor to protect their City they are the most heavily armored of our three classes their designs are most often comprised of a tight fitting rubber suit layered with heavy metal plates with large pauldron likee shoulder pads and sweeping visors inspired by medieval and Renaissance armor large armored thigh pads and boots are also a common trade of the
• 07:30 - 08:00 Titans the badge they wear at their side the symbol of their achievements lineage or Allegiance much like the tabards worn by Knights throughout history okay so we had a distinct look for each of our classes we were happy with but we still had a lot of questions to answer what did different versions of these classes look like what parts would change and what components were key to the class identity how do we break up these awesome designs
• 08:00 - 08:30 to fit into an investment and customization system we generated countless drawings to help us answer these questions and more and we use these drawings to set targets for our customization system but an old colleague of mine used to say designing games is as much about deciding what you can't do as what you can setting limitations was just as important as setting targets for example many of these designs called for great deal of
• 08:30 - 09:00 overlapping pieces which was technically very challenging we did some tests in this area but in the end we felt it was significantly limiting the artists so we just simplified it and moved on we've done customization before with some success with every iteration we've gotten a little more complex and a little more expansive naturally this is where we started could we build off what we had created for the Halo franchise in reach we broke our characters into
• 09:00 - 09:30 helmets chest armor arm accessories and thigh packs this was not unlike our plans for destiny different pieces of gear for different parts of the body the problem however was the scope the sheer amount of content we had to produce for Destiny meant we needed to be thinking outside the box to approach this problem from a different angle during the course of building helmets for reach we created special add-ons intended to add extra variation
• 09:30 - 10:00 to each helmet this worked pretty much as we expected but a funny thing happened we accidentally put the add-ons from one helmet onto another and surprisingly it worked we found that a lot of these add-ons could just be zbuffer into the base helmets to quickly create new variants this got me wondering if we could create an entire character using that approach so I started mocking up a system in Maya building pieces and layering them on top of each other until
• 10:00 - 10:30 I ended up with what you see here 25 or so pieces used to build 12 different helmets using these pieces in different colors I was able to create a variety of helmets very quickly adding in textures and decals in Photoshop I created this proof of concept image this is what we presented to our tools and Graphics Engineers to kick off our early discussions going into these discussions we had several goals in mind for our new system
• 10:30 - 11:00 we wanted to save as much time as possible in both creation and maintenance of our content to have a base of common assets that we weren't rebuilding every time we had to make a new piece of gear how many artists find themselves modeling basically the same pair of boots 10 times throughout the course of a project if they're smart about it they reuse their High poly and kit bash some of their previous assets together but even then they're generally still ripping new maps laying out new UVS and retexturing things countless times we needed to be more efficient than
• 11:00 - 11:30 that that meant the ability to build a shoulder pad one time and use it freely as many times as it was needed in the game without additional work we also wanted to remain as flexible as possible if near the end of development we realized the same shoulder pad was over budget or didn't deform properly or people just didn't like it we needed to be able to fix that piece once and not open every piece of gear that shoulder pad was used in we knew that this was a franchise we
• 11:30 - 12:00 intended to grow and develop over the next decade the system needed to be expandable something we could build on as we added content in the forms of DLC expansions and sequels we knew we couldn't afford to rebuild everything for every release we were spanning console Generations which meant we needed the system to be flexible enough to take advantage of the higher poly and texture budgets that would be afforded by the new generation of hardware and we knew that even within the same generation of consoles Tech improves tools improve artists improve
• 12:00 - 12:30 it was important to be able to address old content quickly and easily down the road from our design team the desire was for player characters to pick up different pieces of gear what we call arrangements for each slot and swap them throughout the course of the game when this happened we wanted the player to notice a significant change to their appearance because of this artists needed to be able to change assets significantly without having to worry about breaking skinning or affecting animation in short we needed the ability
• 12:30 - 13:00 to create something that looked completely different with little to no wake on other teams so after some lengthy discussions with some very smart people we had a plan we needed to stop thinking of our characters like this and start thinking of them like this the idea was to break our character designs down into their most base elements things like belt buckles
• 13:00 - 13:30 shoulder pads visors and straps would all be modeled and textured as self-contained individual assets so building this became much more about building this these individual pieces what we call bits would be stored as separate files and assembled by an artist to create a new arrangement of arms legs helmets Etc if the design for a particular Arrangement called for a buck rather than building a new Buckle the
• 13:30 - 14:00 artist would simply load an existing Buckle bit and place it where he or she needed with the system handling many of the technical aspects behind the scenes to control the material properties of each bit we created dyes the die assigned to a bit controls its color specular qualities detail Maps Etc we decided each arrangement would have a die for armor for cloth and what we call suit which is generally or
• 14:00 - 14:30 rubber each bit within an arrangement will be flagged as one of these types and the material properties set accordingly so we had a plan we knew how we were going to break our characters down now we just needed a way to put them back together enter a mashup mashup is our custom character assembler it allows for any bit created by any artist to be brought in placed rotated scaled to fit the arrangement as needed all without affecting the original bit geometry it really is the
• 14:30 - 15:00 heart of our player character building process so this is an example of an artist working in mashup we start each arrangement by bringing in a core to build off of there's nothing particularly special about the core but it is the foundation of the arrangement much like the foundation of a house house dictates the basic layout
• 15:00 - 15:30 of the structure built on top of it the artist then begins loading in bits and placing them where needed each arrangement has a set of predetermined points to Aid in placement things like front back left upper arm right upper arm Etc bits can be set to default to a certain point just to make things a little easier and faster for the artists as you can see the artist can scale rotate or mirror the bit in a variet of ways to accommodate the needs
• 15:30 - 16:00 of the arrangement there's no hard limit to the number of bits an arrangement can be made of as the bits are combined at export time to save performance as long as the overall Arrangement still fits and its a lotted memory and performance budgets we're all good after we've established the basic geometry of the helmet there are a number of properties that can be set on each bit to alter its function or appearance
• 16:00 - 16:30 we can set what type of material the bit takes from our die system we can set whether the bit draws in first person or not we can set whether the bit excuse me what platforms the bit is drawn on we can also override the bit's default Shader allowing for texture and Shader variation unique to that Arrangement and we can even add vertex animation to give the bit sway or movement after properties are set we need to skin the the bits they can be
• 16:30 - 17:00 quickly rigid bound or skinned using our custom one button skinning tool optionally the bit can be custom skin by one of our riggers and the skin weight stored within the bit and applied the next time an artist brings it in now this video is sped up to about five times normal speed and I'm not going to lie to you I've built this helmet before but it comes in in about 3 minutes and 30 seconds so what you're looking at is a new piece of helmet content in about 20 minutes
• 17:00 - 17:30 so the other tool you saw on that video is Gear Manager Gear Manager allows us to browse through Arrangements bits and Bobs a bob or a bunch of bits is essentially a group of bits stored as a preset it allows artists to put together pieces that are more complex than a single bit but simpler than a full Arrangement a pair of boots including straps and armor might be a good example of this
• 17:30 - 18:00 most importantly Gear Manager gives us a variety of methods to find useful bits the most obvious method being to Simply scroll through pages of thumbnails looking for something interesting but there's much more to it than that I can filter and search bits by name by folder by class or by artist assigned keyword I can also Mark bits as favorites to easily find them later for example I can search for metal bolts limit that to a certain class
• 18:00 - 18:30 quickly scrub through thumbnails and even rotate and zoom a preview of the model all in Gear Manager once I find one I'm interested in I simply double click to send it to mashup going back to our Lego analogy how many hours did you spend paing through that giant container of Legos looking for that perfect piece Gear Manager allows artists to focus on building their Arrangements rather than searching for pieces
• 18:30 - 19:00 cloth setup is generally the last step of our process we use Havoc cloth and a custom wizard likee interface that allows artists to define cloth regions without delving too deeply into technical setup artists pick the type of cloth gear they are making the type of fabric and The Wizard guides them through the selecting the appropriate verts for skinned simulated and anchored regions in the end once it's been imported into game it looks something like
• 19:00 - 19:30 this so now that we've built our helmet we need to make sure it conforms to the geometry and texture budgets we've allowed for it because there's the potential for every player in a given space to be wearing completely unique gear we decided to assign each player their own texture and polygon budgets memory within a mission could then be set aside based on the number of players it required more than that since each player can swap out different parts of
• 19:30 - 20:00 their character we had to further break those budgets down into an amount per slot these were then tailored to best fit each class for example a Titans arm slot allows for a 1024 x 512 texture on the PlayStation 3 and Xbox 360 and roughly 170k memory for geometry a warlocks arm slot which covers less real estate allows for a 512 x 512 texture and 110 K memory for
• 20:00 - 20:30 geometry however the budget for all the combined gear of the Titan and the Warlock is the same okay understanding that let's take a look at this Hunter's chest Arrangement shown in our texture plate preview on the left is the empty plate the green area is the texture space I have to work with for this chest Arrangement on the right is the built plate the engine takes each bit's textures and UVS and programmatically arranges them on the plate anything that
• 20:30 - 21:00 doesn't fit is displayed in the red area on the right side of the preview so here you can see I have three bits not fitting on the plate at this point I have a couple of options I can remove some bits to make more room on the texture I can swap out some of the larger resolution ones for smaller or I can change the bits texture size generally the first two options are my best bet if I change the bit's texture size it's going to affect every Arrangement that bit is used in so
• 21:00 - 21:30 increasing or decreasing resolution of a bit is handled with extreme care in this case though the orange highlighted bit on the left was using an inappropriate amount of texture resolution relative to its physical size so I reduced it down and that change gets propagated to any Arrangement that bit is used in and it allows my three remaining bits to fit nicely on the plate so that gives us a handle on our texture budgets now let's take a look at our geometry
• 21:30 - 22:00 here you see the geometry memory profile for our hunter chest unlike the texture memory each instance of a bit has to be counted so each instance is shown along with a breakdown of cost since many things like UV borders hard edges and Shader assignments can affect the amount of memory associated with the geometry we try to give the Artist as much information as possible if I'm over budget my options are similar to that of my texture budget pass I can remove some bits swap them
• 22:00 - 22:30 out for cheaper ones or I can open one up and make some changes to the actual geometry many times we will find that something simple like a bit containing a large number of unwelded UVS or unnecessary Edge Loops in which case it's a simple matter of right- clicking the bit selecting edit making my changes and saving with that I've just made an optimization to every Arrangement that bit is used in now that I've got my helmet built and budget I'm done right
• 22:30 - 23:00 wrong part of player customization means allowing for both genders so now we have to look at building the female version of our helmet we initially experimented with taking a fully built piece of male gear and using a combination of wrap deformers and blend shapes to create a semi-automated process for building female gear there were several problems with this approach first since it combined all of the bits into a single piece of geometry to then be conformed to the female
• 23:00 - 23:30 physique it lost a lot of the flexibility and benefits of the system second to meet our high quality bar artists were spending longer and longer massaging the geometry by hand rather than relying on the automated mesh so after some careful consideration we made a Direction Change we decided that with a few modifications and additional tools our current process was quick enough and strong enough to use for both male and female Arrangements we added the ability for a bit to have a
• 23:30 - 24:00 female version this is particularly useful for things like chest plates and pants where we want geometry that more closely conforms to the feminine proportions however the majority of things Amo packs and the like just needed to be slightly scaled down and moved into position this allowed us to use the vast majority of the content we built for both males and females now if you noticed on the texture preview slide all of our bits are textured for the most most part is GRA scale
• 24:00 - 24:30 images this is to allow for the additional customization layer brought in by our die system before I get into that I want to talk a little bit bit about the textures associated with a bit like most modern 3D assets our primary maps are diffuse and normal Maps but for bits we also have what we call a gear stack texture here's a sample gear stack texture for one of our hunter capes pretty isn't it what the gear stack is is three
• 24:30 - 25:00 different black and white masks stacked into a single RGB texture to conserve memory we've evolved these masks to be used in a number of different ways but the general usage goes like this the blue Channel holds a fringe map or Alpha test map this is primarily used on cloth bits to create tatters and holes hence the name Fringe the green Channel stores our specular roughness map which works in conjunction with the assigned spec material to control specular r Ness and intensity of the
• 25:00 - 25:30 surface and finally in the red Channel we have our scratch mask the scratch mask is used to control what areas are affected by the die assigned to the bit we call it a d but it controls significantly more than just color areas that are white in the scratch mask get affected by not only the D's color using an assigned blend method but also the D's diffuse and normal detail Maps as well as the specular material response speaking of which let's talk a bit about
• 25:30 - 26:00 how we handle specular our spec response is comprised of three elements the gear specs the gears specular roughness mask as we saw on the previous slide and the D specular lobe and tint the specular lobe controls three aspects of the material spec response the shape of the Highlight the brightness of the highlight and the fall off of the diffuse light instead of using a standard Fong model we map a content driven gradient ramp to the surface to define the
• 26:00 - 26:30 appearance and shape of the specular highlight the ramp textures shown at the bottom of the image are applied to the sphere above beginning at the point of reflection and extending 180° around the mesh the outlined Orange area demonstrates how the ramp textures correspond to the Highlight they produce the specular tint is also a content-driven gradient ramp and it controls four aspects of how specular will render on the surface the color of the specular highlight how the diffused
• 26:30 - 27:00 texture influences the specular response the brightness of the Highlight based on angle and diffuse brightness scaling unlike the lobe ramp which Maps based on the position of a light source the tint ramp maps to the surface based on viewing angle the ramp extends from the surface perpendicular to the view 90° to the glancing Edge this is what allows us to change specular color based on angle of view again the orange outline demonstrate how these ramps get applied to the Spheres
• 27:00 - 27:30 above the tent also sets how the specular highlight is modified by the underlying diffuse these spheres above are rendered in a debug spec light only mode the one on the left allows the diffuse texture of the sphere to modify the specular color and brightness the middle image modifies the specular response based on the diffuse luminance so only adjusting the brightness and the third white Overcoat image ignores the fuse entirely in regards to specular
• 27:30 - 28:00 response put this all together and it looks like this lobe multiplied by tent plus diffuse however this image above assumes a pure white specular roughness mask the spec roughness mask modifies the spec result generally white pixels in The Mask produce a bright sharp highlight and dark pixels produce a dim broad highlight this combination of elements gives
• 28:00 - 28:30 artists a huge variety of interesting spec responses to work with when you combine this with the additional elements in the die it looks like this as you can see this allows us to apply a range of materials from polished chrome to worn leather to Kevlar to tattered cloth all on the same base texture and geometry all of these are materials we can quickly and easily apply to any of our play
• 28:30 - 29:00 gear in addition to the diet areas of gear we also have controls for the other side of the scratch mask the areas that are painted black can be assigned their own specular properties as well as detail diffuse and normal Maps this helps us achieve things like matte paint being Stripped Away from shiny metal or a shiny metal buckle on a leather collar each arrangement gets assigned three different dyes a d for suit for armor and for cloth each D has both a
• 29:00 - 29:30 primary and secondary color associated with it these parameters are set on a per bit basis throughout the arrangement the dies assigned to an arrangement helped to control the color palette homogenize all the bits into a cohesive piece and allow us to form larger graphic shapes within the character the initial design of the customization system called for each player to have three dies assigned
• 29:30 - 30:00 across their entire character this was based on the artist's desire to avoid every color of the rainbows showing up the idea was that players would pick up different die items that would fit into the three slots armor cloth and suit these dies were then applied to the arrangement across the entire character unfortunately this was a case where the design needs required us to adapt our artistic plan
• 30:00 - 30:30 what we discovered was that when a player picked up a new piece of gear even though we were changing the silhouette fairly significantly the visual impact wasn't enough to produce the emotional impact we wanted from the player in other words the player just didn't feel as excited about a new piece of gear as we wanted them to because we were only changing one of the visual axes of the character the human eye reads changes in color better than it reads changes in shape and changes in both color and shape are
• 30:30 - 31:00 necessary for Maximum Impact with this it was decided that each tier of player gear would be assigned a different color so that it was obvious to the player when they picked up a piece of gear that was of higher level the challenge here was to create color pallets that were distinct enough individually but worked well together from one tier to the next here we have our initial passes at creating these color palettes we used simple block representations of our classes to assess
• 31:00 - 31:30 color compatibility and distribution these pallets have been continuously adjusted in engine to better communicate to the player using this color and shape approach proved to be pretty successful in creating bold enough changes to meet design needs while still feeling good to the artists so after building all these fantastic assets we needed the ability to quickly View and manage these
• 31:30 - 32:00 arrangements and die assignments in game so we created mantini with mantini we can quickly select the designer created investment item that a player would pick up in game and set the arrangement and the default dies assigned to it all all within the same UI I can quickly change and browse arrangements and dies on the fly as well as check both male and female versions of the arrangement this proved valuable for checking various arrangements for color compatibility as well as finding
• 32:00 - 32:30 issues such as clipping or missing pieces using mantini we put together this video to show off some of our very early work in progress content [Music]
• 32:30 - 33:00 so that's a pretty solid overview of the
• 33:00 - 33:30 system but I'd be lying if I told you it all went perfectly let's be honest it never does we faced a number of challenges throughout this process the system is fairly abstract and complex Destiny has been heavily concept driven and both the art director and the concept team are far far more comfortable with designs of full characters rather than a modular system though we tried as best we could to create modular pieces that worked well together many of the designs LED
• 33:30 - 34:00 themselves to Unique specialized bits as a result I feel like we've still only scratched the surface of what we can do with the system content creation got too far ahead of tools and Engineering work for much of the production cycle the tools were very much a work in progress missing features and stability issues made working with the system early on somewhat of a nightmare this caused a lot of frustration and slowdowns for the content creators and finally we lost some of the joy of
• 34:00 - 34:30 building characters to be honest this is a problem with any customization system I've worked with it's difficult to maintain a sense of creative ownership and excitement for artists when they're building small parts of a character rather than the full thing since geometry and textures needed to stay modular it kept the artists from really being able to handcraft assets to the level they desired okay so what went right overall the system itself Works quite well at this stage a new and unique
• 34:30 - 35:00 looking Arrangement can be created in a very short period of time for something like a helmet that may take as little as an hour from start to endgame we've also seen the system get better and better the more content we add to it this bodess extremely well for a franchise we plan to carry on for a number of years the flexibility the system provided us has proved invaluable in fixing bugs and reacting to designer and art director feedback don't like the color of something change
• 35:00 - 35:30 the D that sleeve not deforming properly fix it in one place and it's fixed everywhere this fits in extremely well with Bungie's approach of Rapid iteration lastly the system helped to unify content with so many different artists working on player characters the system proved a fantastic tool for helping to set the ground rules to bridge different artist Styles since artists were constantly seeing and working with each other's cont content they quickly picked up on the design language and style we were
• 35:30 - 36:00 striving for the additional layer of the die system assured color and spec responses were also unified so an entirely new type of game an entirely new engine and tool set and a largely new team set the challenge for us and we learned a ton how to best work with the system what tools and features were missing we also learned a lot about our strengths and weaknesses as a team and where we can do things smart harder and more efficiently it's been a tough road but
• 36:00 - 36:30 we've set ourselves up incredibly well for the future and I'm very pleased with that so now that you have an idea of the system and its challenges and successes let's take a look at what happens when you put it all together these are just a few of the player characters you could see when you're playing through Destiny this fall [Music]
• 36:30 - 37:00 let [Music]
• 37:00 - 37:30 [Music] okay so where do we go from here there
• 37:30 - 38:00 were key things we missed out on opportunities to make the system stronger that we didn't have time or resources to take advantage of for this project first and foremost on our list is a more advanced decal system decals are a great way to allow players to customize their look games like Forza have built entire economies on player generated decals our current system requires placement and geometry to be created by an artist specific specific to each arrangement a system allowing for
• 38:00 - 38:30 projected decals based on both Artist as well as user generated content would be a huge tool in allowing players to express themselves developing for four platforms simultaneously simultaneously is a difficult task for any developer to make content more efficiently we need better tools to allow artists to focus on what they do best making amazing art and not having to spend time dealing with separate budget and performance goals for the individual
• 38:30 - 39:00 platforms we would also benefit from an expanded die system additional features in the DI to control things like environment Maps adjust wear levels and blend in multiple sets of detail Maps would greatly improve the system something we've strongly desired is the ability to add an accent color that appears in small controlled amounts to give an arrangement that extra little pop in conclusion Destiny is a hugely ambitious game and our player customization system is no different
• 39:00 - 39:30 like all ambitious efforts there were challenges along the way overall I believe we've achieved something completely unique to the shooter genre and developed a system that will only grow in power and flexibility as we continue to create content and expand the franchise so obviously a system like this takes a bunch of people pouring their Heart and souls into it and most of them couldn't be here today but I have to say thank you very much to the team back at Bungie for allowing me to come here and share this with you that's
• 39:30 - 40:00 it so if anybody has any questions I think they want it you use the microphones but they're not going to turn them on pretend it's on and hopefully you can hear me oh there we go does this system extend beyond the character to things like weapons or World dos and why or why not it does uh we we're trying to use
• 40:00 - 40:30 the system as much as possible um we do use it on our hard surface assets as well like weapons and players ships and that sort of stuff uh when you end up changing a bit how long does it take to propagate that across all the helmets that might use that bit like you have to rebuild all those assets I imagine um the the bits are loaded uh when you um bring the arrangement
• 40:30 - 41:00 into Maya um so from a content creator standpoint it happens instantaneously as soon as it's checked in and you get latest you load up that Arrangement the new the new bit is in there but then when you make the like when you go to use it in the game so for the individual artist they will re they will reexport the geometry and it'll come in game uh other than that we do nightly builds so for people on the other side of the team we'll see it when they get the next content build
• 41:00 - 41:30 okay um so my question is will we ever see tiger man DLC and if not um how does your team go about deciding something doesn't uh fit the world you're building sorry what was that last part how does your team decide um uh something that you've been uh prototyping doesn't fit the art style you're uh taking the game in so I guess I'm supposed to repeat the questions which I totally forgot the first one um question the first part was will we see tiger man in DLC uh the
• 41:30 - 42:00 answer to that is you never know um if if we do you can guarantee I will be using tiger man first before you guys so um second part uh how do we decide if something um doesn't fit the art style um Bungie as I said is very much about iteration we we put things in game and we try to do it in the smartest way possible but if an asset doesn't fit within the art style or within the game play we cut it thank you uh so you have a lot of you
• 42:00 - 42:30 know mix and match hard armor pieces like an upper chest plate and a lower chest plate um how do you make sure that you know all the different possible combinations don't end up self- intersecting and doing bad things and so on um mostly just uh quick and easy tools for checking animations we have um in Maya animation previews so I can be working on an arrangement and hit a single button to see that Arrangement start moving through some base poses to
• 42:30 - 43:00 test for clipping um other than that we have uh we have another program which I haven't shown called lightbox which goes through all of our content create screenshots through various poses that we can then check for any issues I was more concerned about like the combinations of them you know could people mix and match things that so sorry again the question is how do we uh handle clipping of the various combinations of Arrangements um this was something we we decided early on we were going to have to deal with a
• 43:00 - 43:30 little more clipping than than we were used to in previous games I think any customizable game is going to have to accept a little bit of that um we don't have a good way of prohibiting certain things from clipping into other things we just use our eyes and try and best guess is that okay is that shippable is that not let's go back over here um so as in the start of the game you actually choose a a character race um did that end actually affecting the way that you're building gear like did you have
• 43:30 - 44:00 to actually build gear differently over different races or um no the race is sort of a separate part of our character building process although it actually uses the exact same system um uh the the exos in particular are a perfect example of of how we can use the the system for that awesome thank you so this is a really powerful tool for your side of the fence um how does
• 44:00 - 44:30 this level of customization and content generation trickle down to the player's ability to customize and leverage the amount of flexibility that you have um the question was uh this is a really powerful system for us as content creators how does that affect the players and how do they make use of it um right now for the most part it affects them in the fact that we can build a lot more content than we could otherwise we can build a lot more armor
• 44:30 - 45:00 sets a lot more um pieces of gear for them to choose from um pushing forward exploring the ideas of players being able to create their own arrangements and stuff is something we've definitely thought about and you know I would personally love to explore but it's a long ways out um how does the bit system affect the concept pipeline um I does uh do the concept artists then create more of a kind of key uh keynote versions of everything and then the variations on the individual bits are more on the 3D
• 45:00 - 45:30 artist and create on the fly so the question was um how does the bit system affect the concept development um and how the concept artists work um it's a great question uh and you're exactly right um the concept artists create what we call theme so they have a set that is like here is what the sort of uh the spirit of this set and then the variations are all built by the indiv individual 3D artists for the most part um we try we very much a Bungie don't like to limit our concept artists but we
• 45:30 - 46:00 also want to educate them on the system so um they start to know like what we can do and you know what things are easier and faster for us um but still they have kind of free free range to explore their ideas thank you um yeah I was wondering about upgrades and if there was a way that you guys had like rules or or some way of defining like what an upgrade would look like since like color doesn't necessarily denote hierarchy or there
• 46:00 - 46:30 other things that used um so the question is about upgrades um and honestly I'm uh I'm not a a designer so the investment designers choose sort of the upgrade paths and what pieces they want to use for those um that stuff since they're constantly constantly iterating a lot of times um we you know uh we try and put in the best pieces we think possible for things um but they're much more focused on the
• 46:30 - 47:00 stats and the abilities and that sort of stuff so I don't know if that answers a question but so you didn't have any like uh ways of defining like what those upgrades would look like necessarily um it the way the system sort of played out we very flexible and reactive so if they say hey we have this thing and we're going to do an upgrade to it then we can quickly go in and say all right let's add a few more pieces make this thing a little AR more Burly and a little cooler looking Okay cool so it's it's more like they establish first and then we kind of follow what they what they're looking for okay cool
• 47:00 - 47:30 thanks uh so with the bits um you said that it would go back and change other uh bits that were already put into uh other sets so uh was there any instances where a bit being changed in a newer set changed an old set for the worse so the question was uh since we can change a bit and it affects other arrangements where there instances where this was bad where a bit was changed and it broke something the answer is
• 47:30 - 48:00 definitely yes um we had some interesting cases where someone would the most dangerous aspect is um the texture resolution someone would decide hey this bit's too low res and so they'd double it and then suddenly we have you know 20 pieces of gear that aren't importing because their texture budgets blown um so that was something that we you know through a certain part of the production we dealt with and and then artists just kind of got smarter and like better about knowing what they can
• 48:00 - 48:30 and can't do and and also better tools at checking the the content afterwards thank you um so I think you said that the uh bits get combined at export time and then so I'm assuming the arrangements get stored and loaded by the engines that every Arrangement is a separate piece that's rendered by the engine is that correct or that's correct um sorry go ahead oh yeah I just wanted to check like isn't that uh more expensive than rendering a traditional
• 48:30 - 49:00 character perhaps so so there more drill calls associated with rendering each arrangement uh so the question is is um you know does the engine render each arrangement individually and is that more expensive uh than sort of a traditional um full character Arrangement or full character setup um the answer is yes it is more expensive uh we've done a lot of things to optimize the um the GE Shader process and um you know the skinning and all of
• 49:00 - 49:30 that to try and get that as cheap as possible but definitely the characters the player characters are more expensive than a lot than say civilians or combatants or anything like that can I ask another quick question just um I think you said you had a lot of problems with wrap wraps and uh blend shapes can you go but you didn't sort of elaborate what the sort of problems were with you know cuz I would have thought that would give you close fitting cloth if you could deform layers of cloth over over deformable shape um so the question was about uh I I mean I think using the
• 49:30 - 50:00 wraps and blend shap wrap deformers and blend shapes to create female gear and whether not why that wasn't successful um and you're right like for things like cloth and soft Organics it worked very well but as you can see we have a lot of hard surface shapes and things like cylinders that you know would get stretched and shapes would get deformed and the main thing for us was that in order to really use that system we had to combine everything together and that just wasn't efficient okay thank you I
• 50:00 - 50:30 had a question about the color customization if I understand correctly the originally the player could choose three colors for their character and then you guys changed the system to colorize the gear so it would represent the different scale upgrades um so does that mean that in the current design there's no player color customization for their character so the question is um we initially had it so that players could choose three colors and then we moved to a system where that uh where we establish the colors to to communicate
• 50:30 - 51:00 the tier does that mean players Can't customize their color um so uh no we we we definitely established what we call default dies for system for things and um the the idea is that players uh will be able to customize their color choices later on they can take their upgraded gear and then tweak it from its current value basically um so I can't go into too much detail about this because it's actually uh something that's still being worked on um but uh as I said the idea is that
• 51:00 - 51:30 players will have some control for sure okay uh hi I was just wondering about the the merging process you talked about so when the bits are merged together um is there at that time is there a lot of inter penetrating surfaces or do you do any kind of Ray casting to make a new mesh or uh if if one if one Arrangement overlaps another Arrangement do you make a bake that has like nice AO for the under the overlapping or do you just try not to make them over I guess that's two
• 51:30 - 52:00 questions um so the the sorry the the first part of the question again um like the helmet you showed with all the bids together um if it generates in inter penetrating surfaces is there a bake process that makes a new mesh or it's just okay to have the inter penetration um so the question was is does the do the bits have interpenetrating surfaces is that okay do we bake things down um do interpenetrate and uh we rely primarily on screen space AO to kind of seal
• 52:00 - 52:30 everything together and make it feel like it's a nice um you know cohesive piece um and then uh the second part sorry well that kind of answers it you know like a gun all right thanks Justin quinby hi Scott great talk uh we work together at turbine uh so I'm curious about the production side of things in terms of dashboarding tools or or how do you how does the production side of the house track you know you
• 52:30 - 53:00 mentioned oh suddenly someone changes texture density and blows uh the the budgets for stuff what kind of tools do you have on the production side of the house to uh keep track of all the craziness that you can do giving giv this flexible tool set um so the question is is what what kind of tools do we have from the production side of things to keep keep track of all these changes and and all the things that can be done um honestly uh nothing terribly fancy um we have fantastic producers really that are just very on top of
• 53:00 - 53:30 their game um we don't as a 3D team we don't microt track stuff too much um as our tasks tend to be you know this this armor set will take four weeks or three weeks or whatever um and then we just try and remain as react as as reactive as possible the good thing is is that because the system is so fast and flexible when we do run into those issues it's it doesn't take much time to fix it hi uh first I want to say great talk um
• 53:30 - 54:00 I was wondering uh if the current system for creating uh like different pieces through bits and stuff allows for the addition of like normal map details such as adding like screws or uh you know different like vents or valves on like a helmet or something or if that's part of the decal system you guys are looking towards in the future so the question was is if if the the current system allows for adding uh sort of normal map details like screws and things like that um uh actually the um it does in a
• 54:00 - 54:30 different sort of way so we have uh we have a layer what we of what we call detailed geometry um which is set within set as a bit uh so certain bits can be flagged as detail geometry so they're high detail high frequency type of bits that then fade out quickly at distance or don't get exported to sort of the older platforms kind of thing so we do a lot of that with those thank you um I uh one quick question I had cuz the system it actually reminded me a lot of Borderland 2's gun system where they
• 54:30 - 55:00 built completely modular guns and so each player who started up their game was able to have like their own unique sniper rifle um would this system allow for players to have like completely unique randomized helmets or is everything built beforehand so the question was um does the system allow for players to have completely unique randomized helmets um or pieces of gear um the answer is no actually and and that was a very cont decision on our part uh we wanted artists to still maintain control of the overall design
• 55:00 - 55:30 um we wanted to try and make it as fast as possible for them to create something cool but we didn't want it we always wanted it to feel like it was a conscious decision that we put this piece with this piece thank you all right any more questions you guys are free to go thank you very much e