Open main menu

UESPWiki β

General:Illustrating Tamriel: Creating Key Art for The Elder Scrolls Online

< General: Interviews
Illustrating Tamriel: Creating Key Art for The Elder Scrolls Online
GEN-misc-ESO Key Art GDC 2019 Screenshot.jpg
GEN-misc-ESO Key Art GDC 2019 YouTube.jpg
(link)
Medium/Format Live Presentation
Date March 2019
Interviewee(s) Lucas Slominski
Location Moscone Center
Hosted By Game Developers Conference

My name is Lucas Slominski and I'm a senior concept artist for ZeniMax Online Studios. Hold on one second, I jumped the gun on myself here. Okay. Yeah, I'm senior concept artist for ZeniMax Online Studios where my main responsibility is producing key art illustrations for the Elder Scrolls Online. And just to give you a bit of background on myself before we get started, I originally went to college for illustration at Syracuse University in Upstate New York. [Muffled, distant cheering] A lot of Syracuse fans next door. And although Syracuse is a great school, their traditional media curriculum didn't really prepare me for a career as a concept artist, which had always been my primary career goal. So after Syracuse, I decided to go to grad school at the Savannah College of Art and Design where I focused more specifically on 3D art for games. And shortly after graduating from SCAD, I was hired at Turbine in Boston where I worked on the Lord of the Rings Online and a short-lived DC Comics MOBA called Infinite Crisis. And then in 2015, I moved to Baltimore to work for ZeniMax Online Studios where I've been painting the key art illustrations for the Elder Scrolls Online ever since. And I imagine a lot of you are familiar with the game, but in case you're not, I'm just going to give a bit of background on the Elder Scrolls Online, or "ESO" for short. Our game is an MMO set in the Elder Scrolls universe, taking place about a thousand years before the events of Skyrim. But unlike preceding Elder Scrolls titles, ESO is not developed by Bethesda, but by one of its sister companies, ZeniMax Online Studios, or as we call it on the dev team, ZOS. The game launched on PC in 2014 and on Xbox and PlayStation in 2015. It currently has around a million monthly active players and we continue to support that player population with a regular release cadence of four major content updates per year. And as part of the marketing campaign for each update, ZOS tasks me with producing a highly detailed hand-painted digital illustration in-house, and of course that's what we're here to talk about. So my goal for today is to give you a peek behind the curtain at ZOS, showing the process I use to create these illustrations and the various competing influences other than me that ultimately shape the outcome of the final artwork. So more specifically, first I'm going to try to clarify exactly what I mean when I refer to "key art" for ESO, plus how it's used to support the game, and the main parameters we use to evaluate the success of completed marketing illustrations. From there, I'm going to show some examples of how I try to balance the sometimes contradictory demands of the various parties that have a stake in the key art, and once we understand how they serve to limit my options when designing compositions, I'm going to dig really deep into the details of some of the strategies that I use for visually crafting narrative within single-image illustrations. And then finally, I'll try to show how I put all those ideas into practice by walking through the full production process for one of my completed illustrations. It's an ambitious itinerary and we're already on the clock, so let's get started.

"Key art" is a term that can mean a lot of different things in different contexts, so rather than get stuck in the weeds on terminology, I wanted to find key arts solely according to the way we use the term on ESO. When we talk about key art at ZOS, we're specifically referring to the illustrations that we use to market new content releases for the Elder Scrolls Online. They're single images intended to be the primary visual representation of the content release they each accompany. So basically, you can think of key art as the rough equivalent of box art, but for a digital game. The key art is featured on the ESO website and gets used in ad banners across the internet and they're also released as desktop backgrounds and promotional posters at conventions and other fan events. And just to be clear, I should mention that the marketing campaign for ESO's big annual expansion in the second quarter of each year also includes a cinematic trailer, a logo refresh, and a CG key art banner which are all built using CG, 3D assets. All that stuff is outsourced to creative agencies, which means I'm not involved in their production, so for the purposes of this presentation, I'm not going to be including any of these within our definition of key art, and we're just going to be pretending that these things don't even exist. Okay, so this is what the key art production process looks like at the broadest level. We start with 2-3 weeks of planning and development, where I research the story and iterate on several rounds of comps - short for compositional sketches - before getting the greenlight to go ahead with a particular composition. I then move on to the main production phase in earnest, which usually takes about six weeks. So the whole process takes about eight weeks in total, which is a really long time spend working on a single illustration and nobody knows that more than I do. However, my art director prefers that I focus on realism and detail rather than speed and he likes me to take a constructive approach to figure production. And because of the variety of potential use-cases or the key art, I also paint it at a very high resolution which allows for smaller areas to be repurposed for different uses in the future. And all of these things contribute to a longer production time frame. And since ZOS invests so much into these images and they feature so prominently in the way the game is presented to the public, the key art sits at the intersection of various competing interests and the stakeholders for those interests each get varying degrees of input into its development. So traditionally, it's more typical for a studio to outsource the production of marketing assets like key art to a freelance illustrator or creative agency and the discrete nature of that vendor-client relationship forces all revision requests through a single point of contact, usually represented by a marketing art director who is typically not the same person that's in charge of the in-house game dev art team. But because I produce the key art in-house as a member of the development team, our art director at ZOS is able to exert a much more direct influence over the final look of the key art, which tends to result in the illustrations more effectively intimating that same character and tone that's shared by the rest of our game's content. But it also means that I'm more directly confronted by the needs of different stakeholders than I would be under the traditional freelance model. So let's diagram those stakeholders and my relationship to them. As the illustrator, I sit in the center of this diagram and the ESO art director represents my primary source of critique and feedback. But he's only one of several members of management to get a say in which of my comps is approved for production. I call this group the executive team and it includes ZOS's studio director, the general manager, and finally the Elder Scrolls brand manager who works at Bethesda and represents their marketing department. Each member of the executive team has de facto veto power over any of my compositions. My art director insulates me from direct communication with the other executives as much as possible so that he represents the single source of directorial feedback for me. Now, in one sense this is really helpful because I'm not faced with the challenge of reconciling contradictory requests after submitting each round of sketches, but on the other hand, when my art director has to relay feedback on behalf of the other executives it can sometimes be tough to determine the essence of the original critique that I really need to respond to. It also makes it impossible for me to advocate for my own concerns directly, so I have to rely solely on my art director's impulse to push back against decisions that I might not agree with. I also get content-related input from the ESO loremaster - and, yes, that is his actual job title. And he ensures that nothing I'm including in the illustration contradicts rules previously established in other parts of the Elder Scrolls universe. So while the left side of this diagram is full of stakeholders telling me what the key art needs to look like, the right side will represent downstream dependencies for the key art, and none of the stakeholders on this side have directorial oversight over my work, but they do each need me to hit particular benchmarks to facilitate their own uses for the artwork. So the creative services team needs the key art to conform to specific technical parameters so that they can process it for various use-cases before handing it off to the web team, which publishes it on the ESO website, along with promotional copy written by ESO community managers who use the key art to help frame the focus of their content, while also sharing the key art in social media posts. Meanwhile, my producer manages and enforces my deadline in concert with the lead concept artist who really functions as my my manager and helps to facilitate my interactions with the other teams. So with all of these parties involved, my task in creating key art sadly isn't as simple as just painting a pretty picture. In fact, in addition to the technical specs required by the key art's eventual use, the content of what I paint is more broadly guided by a series of persistent marketing goals. And we use these as parameters to gauge the success of proposed sketches and they greatly influence how and why I arrive at my final compositions. So, here they are. First off, the primary purpose of the key art is to generate player excitement for new content updates. And, of course, we want to attract players that would be interested in playing the type of game we're [sic] actually made, so it's important that the key art help establish expectations for the specific gameplay experience we're offering. ESO is not a horror game or a comedy, it's a social dungeon questing RPG, and the key art needs to convey a tone of adventure and meaningful character interaction that's representative of what players will experience when they log in. And, of course, the key art needs to highlight the new features and setting of the content update that it accompanies. This seems obvious, but it does have more detailed implications like decisions to clothe the characters in newly-created armor and outfits to reflect the gear that players will be able to collect in the accompanying update. ESO is also extremely story-driven and the key art needs to reflect that by being focused on narrative as well. This is a major point of emphasis for my art director. He's always pushing me to inject as much narrative as possible into the key art, and, as we'll see, that heavily influences the content of the compositions that I develop. And then finally, another important requirement is that our key art feature emotionally resonant player characters. Emphasis on the "player character" part. Bethesda's marketing department insists that the key art for ESO include figures that the player can imagine themselves inhabiting based on the type of experiences offered to them by the gameplay. That means that I can't just depict off-screen story moments occurring between NPCs. A figure representing the player must always be involved in the action, and that demand introduces a whole host of complications that require creative solutions, and we'll be talking about that more in just a minute.

Now, with all that in mind, you can see how planning one of these illustrations is an exercise in problem-solving. Luckily, the demands that I'm talking about rarely contradict one another outright. However, they rarely all serve in the best interest of achieving an elegant composition, which is my primary concern. So the struggle for me becomes finding my way to an aesthetically competent visual solution while also accommodating external requests that severely handicap my ability to do that. And to demonstrate how this works in practice, what I want to do now is highlight three specific challenges that I come up against pretty frequently and which have a significant impact on how these illustrations tend to turn out. And to make it easier to distinguish between them, I'm giving them each a catchy name. So we have the Mandate, the Blackball, and the Chimera. I'm going to explain each of these one-by-one and show examples of how they impact my work. So we'll talk about the mandate first. And it's basically my shorthand to refer to that last parameter I mentioned earlier where I'm required to include player characters in every key art illustration that I produce. This tends to be extremely problematic both compositionally and narratively, because ,if you're going to put an extra character in a scene, that character needs to have a reason to be there. He has to have an important role, because we don't want to give the impression that our players are essentially spectators to the action. Additionally, player characters are deliberately conceived of as blank canvases onto which players can imprint their own stories, so there's nothing specifically unique about raw player characters' identities or backstories that enables viewers to engage with them. And that means that the relevance to the illustration's narrative has to be derived performatively. That is, they have to be doing something. Unfortunately, as with any game type, there's a limit to the types of activities players can actually participate in while playing in MMO. Some of which, I've listed here. And we could depict scenes of players engaging in these activities, but that would violate some of the other items on our list of goals and parameters, right? So, for instance, as a rule, available player activities remain the same for each update, but key art is supposed to be focused on new content, so I can't just do, like, a crafting illustration or a fishing illustration and call it key art. There's nothing exciting about that. And, in part, that's because those activities don't lend themselves to a compelling narrative, and our key art is supposed to be fundamentally focused on the main narrative of an update. The activities that I listed in the last slide may represent ways that the players experience new story content, but they don't represent the unique appeal of that content, and that's something that the key art is supposed to do. Furthermore, the activity depicted in the key art can't be so specific to an individual quest that it ignores the larger role-playing theme of the update. Now, when new systems are introduced in an update, things get a lot easier for me. For example, when we launch a new guild that players can join, then we can specifically depict players engaging in the activity defining that guild. So, the Thieves Guild is obviously a fraternity of thieves and so the key art for our Thieves Guild update could logically focus on player characters committing a heist. Likewise, the Dark Brotherhood is a group of assassins, and so the key art for that update naturally depicted player characters performing an assassination. Similarly, our Morrowind expansion included the new warden player class, so we had an obvious opportunity to showcase player characters utilizing the new class's skill line. And then, finally, our homestead update introduced a new player housing system, so an ideal solution was to show player characters inhabiting a private home. These pieces fulfilled the new content and main narrative parameters because they each more-or-less represented the marketing focus of the update and because they established a new role-playing concept for us to highlight. But without a new system to feature, the marketing push typically focuses on the new playable zone, or dungeon, or story-based theme. And finding ways to meaningfully include player characters in those becomes a real challenge. In these cases, I have to ask myself the question: "What's the difference between a player character and an NPC, visually speaking?" It's a question that doesn't really have a specific answer. For example, one of the two compositions you see here was rejected for a lack of player characters in it. Which one is it? Let's go for some audience participation here. I know these are just unfinished line drawings, but raise your hand if you think the one on the left failed to include player characters. Okay, and then raise your hand if you think it was the one on the right. Alright, so, people who chose the left were correct. But I'm not surprised that the room is torn. It's an unfair question because oftentimes the player character designation is only clear if you already understand the backstory of the quest design, and that's information that most consumers won't have when they first see the image anyway. In short, there is no significant visual distinction between a player character and an NPC, which is kind of the point. So what's the executive team's way of identifying player characters in these sketches? Well, in my experience, they seem to recognize player characters in key art according to the following parameters. They're looking for a cluster of three figures of different races who are typically foreigners to the the location of the new content and preferably outfitted with gear indicative of different player classes. Now, these parameters have never been explicitly defined for me, this is just what I've gleaned over time based on which of my compositions have been rejected and which have been greenlighted. But the hidden cost of having to accommodate the player character mandate is that it forces me to forsake a lot of great off-screen story moments because there's just no way to inject player characters into them.

Okay, that is the Mandate. So now let's talk about the Blackball, which is basically the Mandate's opposite. So, the Blackball represents the kind of seemingly-random blanket prohibitions that executive stakeholders sometimes place on certain compositional elements. One of the more problematic of these is the directive that I can't show main characters that are turned away from the viewer. Now that might not seem like a big deal at first and it's certainly an understandable request given that heroic character backs have become something of a trope in marketing imagery. And, admittedly, avoiding the overused examples shown here is not a bad idea. But there are also many cases where it's compositionally and narratively useful to depict the main character facing away from the viewer, and the reason for that has to do with the limitations of single-image visual narrative. We're going to get deep here, okay? Fundamentally, the spatial link between a figure's face and the object of their attention results in two parallel image planes that face one another. And it's generally impossible to see the fronts of two surfaces at the same time when they're facing in opposite directions. Now, cinema deals with this by cutting between multiple reverse shots. So, one shot shows the character's POV, while a second shot will show the character's reaction to what we just saw. But in key art illustration, we don't have the luxury of multiple POV shots, so oftentimes we're forced to choose between focusing on the character or focusing on what she's looking at. And doing the latter typically means we see the character's back instead of her face. Now, while focusing on the character is certainly the primary conduit for evoking emotion in the viewer, the character's object of interest is oftentimes more powerful in being narratively informative. And, why is that? Well, in its simplest form, you can think of narrative as the presentation of cause and effect. First, there is an inciting incident, and then drama emerges from the way the consequences of that incident play out. For example, if an earthquake causes a tsunami, the predictable consequences of death and destruction are likely to follow. In character-centric illustration, whatever challenge the characters are shown to be confronting is what occupies the role of cause and the character's reaction to it represents the effect. As a rule, an effect is more easily inferred than it's cause. So, it's easier for a viewer to guess at what a character's reaction to an incident will be than it is to guess at the situation that was responsible for producing that reaction. The viewer needs to see evidence of whatever situation a character is facing in order to understand what's at stake and to sympathize with what they're experiencing. So, returning to our tsunami example, if we want to plan an illustration that conveys the essence of the rudimentary narrative we've set up here and we can only focus on one of these story beats, which image should we keep in order to retain as much of the story as possible? I would argue that the first image is more useful. Consider if we were to take that first image away and only see an image of death and destruction. Well, there are any number of situations that may have led to that result, including a tornado, a fire, or some kind of military action. But if we keep only the inciting incident, well, its resulting consequences are usually much more predictable. Conversely, an illustration that only focuses on a character's reaction to a conflict without showing what that conflict is, well, that robs us some very powerful storytelling tools. And since we want to show both the cause and its effect within a single illustration, the problem with the Blackball demand is that it requires me to make two facing planes visible to the viewers simultaneously from a single point of view. The most straightforward way to do that is to view the scene side-on, but that forces the composition to be set up like everything is occurring on a theater stage, which is pretty boring. Of course, it is possible to find other solutions to this restriction, but the Blackball still significantly limits the types of images and stories that I can effectively tackle in ESO key art.

Okay, so that's the Blackball. And just to review, the Mandate forces extra elements into the key art, while the Blackball restricts the options I have at my disposal to incorporate them gracefully. Collectively, they limit my ability to generate clear, concise compositions for new key art, which in turn leads to the third major challenge, which is the Chimera: expanding the scope of a composition to accommodate disparate marketing goals. This one's the killer. The Chimera basically refers to efforts to shoehorn extra compositional elements, such as player characters, into a narrative that wasn't designed to fit them. More often than not, this comes about when stakeholders propose their own solutions to the problems created by the two preceding challenges. And it's by far the biggest source of scope creep I encounter with the key art. So, as an example, let's take a look at the set of comps that I developed for our 2017 third quarter update, the Horns of the Reach. And each comp you see here depicts an off-screen story moment extrapolated from the content of the main quest. Now, while I knowingly omitted player characters from the top two sketches, the bottom two were conceived to allow for certain figures to be interpretable as either player characters or NPCs. In spite of that, the executive team preferred this sketch from the top row. But since it didn't include player characters, they requested that I modify the composition to add them in. Now that was problematic because the ritual that I'm showing take place here doesn't allow for the player to have any active role in the scene. Players don't actually participate in or witness this ritual in-game (it's only mentioned by NPCs) and player characters just wouldn't make sense within the context of the story that this composition was intended to tell. But, due to production time constraints, and because the stakeholders had already endorsed this composition, I had to find a way to fit player characters in there somehow. My art director's solution was to expand the image by framing this composition with player characters sneaking up on the ritual in order to stop it from being performed. So, that caused the scope of the illustration to expand immensely. It also created this strange image-within-an-image effect where the foreground scene of the player characters framing the action that's taking place in the background makes the original scene seem even more vignetted and remote. Not to mention, it more than doubled the total surface area that I had to render and resulted in a lot of after-hours and weekend work to get this thing completed before the deadline. And the real kicker was that, for the final ad campaign, somebody decided to crop the image back down to my original composition anyway for most social media posts. So we are not a fan of that. Get that out of there. There are two relevant takeaways from this example. The first is that adding characters dramatically increases production time. While my producer tries to accommodate me when the scope of an illustration changes, oftentimes the additional work required exceeds the additional time available for me to complete it. So it's in my personal interest to avoid character bloat, but, more importantly, showing more oftentimes doesn't make the composition better. In fact, it may overcrowd the image, complicate the read, and make the picture feel too busy. Now, the image-within-an-image framing device that ended up being used in that last example is worth a little more discussion before we move on, because it's an idea that gets suggested semi-frequently as a way to stretch a single illustration on ESO to include more complex plot elements than it can comfortably hold. And to show what I mean, let's take a look at the selected sketch for the Psijic Order key art that I painted for ESO's Summerset update where we can see all three challenges at play. So, first, the Chimera issue is represented by the picture-in-picture effect created by imagery appearing inside a magical holographic vision at the top of the sketch here. That's because in the corresponding content, the mages of the Psijic Order were spatially removed from the events of the main questline, so I was asked to include this floating magical snowglobe element as a way of symbolically representing the threat that the players would face in the main story. Also, the Mandate came into play, because my original drawing didn't explicitly include player characters, so my art director requested that I restructure the composition to fit a trio of player characters into a horizontal crop format. And finally, the Blackball restricted the options available to me to make those changes, because I couldn't position foreground characters with their backs turned toward the viewer. So, adding the triad of player characters created a very unbalanced image composition, so I was able to successfully argue to get one of those player characters removed. And the final composition ended up being reasonably effective, with the exception of the picture-in-picture motif represented in the center. It's difficult to parse exactly what it is that we're seeing inside that magical sphere, and the fact that its meaning can only be understood symbolically doesn't do the narrative any favors. It was a struggle for me to even get the illustration to this level of readability, and, since then, I've used it as a cautionary example to dissuade similar suggestions for more image-within-an-image compositions. Okay.

So, now that we've covered the major challenges that might limit my options when conceptualizing the key art, I want to show some of the tools that I rely on to maximize the narrative impact of an illustration without overloading and breaking it. As a case study, I'm going to briefly walk through the development process of the key art I created for ESO's Dragon Bones update, which I painted back in December of 2017 and which was released the following March. And, as we'll see, most of the creative work for the key art is frontloaded to the earlier stages, the actual production and rendering of the image is a rather mechanical exercise. So the whole process starts with an initial brief from my art director which essentially amounts to a really basic overview of the new content along with his ideas for particular story moments or important visual elements that he thinks we should highlight in the key art. As I mentioned, one of the things my art director really likes to emphasize is a narrative focus for each illustration we do. It's important to him that each image tells its own story. Now, the Dragon Bones update was a dungeon pack, which, in the cadence of the ESO release schedule, tends to mean that it's relatively thin on story, but the basic gist of it was that a necromancer had discovered the bones of an ancient dragon which he intended to reanimate and use it to kill all of his enemies. The player's goal is to stop him, and in doing so collect cool loot. So maybe not super original when viewed at a high level, but it's definitely enough to get us started. So after talking with my art director I do some more research on my own, which includes reading the design documents generated by the content designers, looking through the concept art that's been done for the update, and, depending on whether it's been built yet, I fly through the play space to get a feel for the environments that players will encounter while playing. So, what do those story elements that I mentioned determine for the key art? Well, right away, I knew that this key art was obviously going to have to showcase a reanimated dragon skeleton. It was the central focus of the dungeon pack and you can't call something "Dragon Bones" and then ignore those bones in the marketing campaign. Also, since it's unusual, even in Tamriel, for a dragon's bones to come back to life, I knew that it would be important to show the player characters' shock and surprise at encountering a giant living skeleton. I also needed to provide some explanation of how, or why, the dragon skeleton has become reanimated, which meant I needed to create a visual link between the necromancer and the dragon so the viewer understands the causal connection between them. And then, finally, in order to convey the enormity of the threat the dragon poses to our heroes, it would be helpful to emphasize the scale of the dragon compared to the size of the human characters. So with those cues as a jumping off point, I started doing some very rough thumbnail sketches, and you can see how these are all really built on the elements that I just talked about, especially, like, the scale difference between the necromancer and the dragon and the player character's reaction to seeing it. The sketch in the lower lefthand corner had the most developed narratives, so I started fleshing that one out first. As we discussed, a major storytelling constraint of key art illustration is that we're confined to telling our story within a single image. And that makes it difficult to convey multiple story beats that take place over time. One of the ways that I sidestep that issue is by using different characters to represent different beats or moments in time. A successful reading of the image is therefore dependent upon constructing the composition so that the viewer's eye notices each of these story beats in the order we want, and we can see that playing out in this first comp here. So, the initial read in this composition is clearly the necromancer silhouetted in front of the dragon skeleton. And based on his pose, we get the sense that he seems to be casting some sort of spell on the bones. Next, we notice the figure in the center foreground who appears to be sneaking up behind the necromancer with the sword drawn, along with that figure we also pick up on the foreground figure on the left who's urgently gesturing towards something on the right side of the image, and that carries our eyes to the figure on the right who seems to be startled by a light glowing from the eye socket of the dragon's skull. We also realize that the central foreground figure seems unaware of the glowing eye lights, suggesting that they've only recently begun glowing sometime after he took up his current position. Now, of course, when we first look at the image, our minds take all this in within a second or two, but the order that we notice this stuff influences our ability to easily grasp the story and learning how to direct the viewer's eyes to move through a composition is an important aspect of storytelling through illustration. The artist has a number of tools at their disposal for directing the viewers gaze - and this isn't really the focus of today's discussion, so I'm not going to go into detail with these. But by playing with relationships between lightness, color, scale, and proximity to the viewer, the artist can establish lines of action along the edges of major shapes, which is what guides the viewer's eyes through the composition. And there's one other important composition planning tool that I think frequently gets overlooked, and that is that in the west we're taught from an early age to process visual data by reading from top-to-bottom and left-to-right. That means that when a viewer approaches a work of art for the first time, their eyes will automatically seek to enter the image from the upper left corner by default. So it's important to take this into account when planning for viewer eye movement, because it gives you a reliable start point where the viewer will enter the maze of your illustration. And just as importantly, if your goal is to have the viewer's eye first enter the image at a different location, then you have to build in a lot of extra visual cues to counteract that top-left corner bias. So with that in mind, when we refer back to our dragon bone sketch, we see that I'm using the dragon's upper arm bone like a highway to lead the eye in from the upper left directly into the center of the image. From there, our eyes will continue to move around the composition following the lines of action created by other major shapes. Our eyes naturally stop to rest at any location where one line of action ends and another begins. And these areas, thus, become the focal points of the image and the location of different story beats. So after taking in the necromancer, the bright high-contrast line created by the dragon's nose channels our eyes down to the character in front where the angle of the sword and the direction of his body - that his body is facing - swings our focus back to the left, encouraging us to engage with the figure positioned there. Next, your eyes pick back up on the high-contrast edge created by the rim lighting of the dragon's forearm and continues from left to right along the upper arm and spinal column. Now, arcing our focus over the top of the image like this is an important step, but you could also say it's a dangerous one (insofar as there's any danger in illustration) because in tracking along the strong rightward descending shape of the dragon's spine, I'm reinforcing our eyes' predisposition to move from left to right and the risk is that the viewer's attention could drop right off the edge of the image, which is obviously not what we want. So that's why it's important to build interesting shapes like the dragon's horn and the rim lighting along the figure's outstretched arm which provide a directional counter-force to bring our focus back toward the center of the image. And at this point we see that we've established a circuit of visual interest which invites the viewer to retread the moments of the story in a clockwise fashion and in so doing encourages them to spend longer engaging with the artwork. And when you spend as much time painting these images as I do, you definitely want to squeeze every second of your engagement that you can out of them.

So not to belabor the point, but quickly going back through that narrative sequentially, this is the full story that we'd be trying to tell if we had more than one image available to do it, right? The necromancer begins casting a spell on the bones, our heroes will sneak into position near the skull, the dragon's eyes light up, some of the heroes notice the eye lights and react, and then finally Dragon wakes up and our heroes become overmatched. Unfortunately, we only have one image at our disposal to tell the story and technically this snapshot has to occur at a single moment in time. So I'd say that what we're seeing in this comp is taking place right about here: somewhere between beats three and four. But of course, the first four beats are all represented to some extent in the comp even though some have already occurred. Because the storybeat of the glowing eye lights is the last stop on the viewer's guided tour through the composition, our eyes interpret it as having occurred more recently than the other beats in the image even though everything is ostensibly occurring at the same moment in time. But based on the way I've structured the composition, I'm providing clues to what has happened in the scene just before the moment we're witnessing and giving the viewer hints of what is likely to happen in the moments to follow. So, in this case, the dragon is going to wake up and these guys are going to be in trouble. In this way, I'm establishing a story with a beginning a middle and an end all within just one image. Now, the feedback that I got from my art director on this sketch was that he wanted the player characters to have more to do than just sneak up on the bad guy, and this is yet another example of the problem created by the player character mandate. Searching for solutions to that problem led me to this composition here which shows our quest party attempting to steal loot from a dungeon. Obviously, the dragon is the first major read here, but the critical element of this story is our heroes' ignorance of the danger they're in. So like this guy is so focused on breaking into the vault that he's oblivious to the massive reanimated skull position right above him, the lookout on the right thinks he might have heard something but he's looking in the wrong direction, while the coiled posture of the figure holding the pickaxe tells us that he's just clued into the real danger from above. And then over on the left we have a confrontation going down between one of our heroes and the necromancer who's being held at sword point. It's clear that our questing party believe they have the situation under control, but the necromancer's glowing staff indicates that he's magically controlling the as-yet-unnoticed dragon, so he's not as helpless as he looks. Now the main storytelling problem with this composition is that the necromancer isn't prominent enough in the image. Using the eye movement analysis we talked about before, the necromancer is one of the last reads and the nature of the relationship between him and the dragon isn't very clear. So, to try to solve that problem, I kept the same stage, but repositioned the actors to seize some control back to the necromancer. So now he's the one in the position of power above our heroes. And notice that the player characters each have different levels of awareness of the dangers facing them. So like the guy closest to the necromancer is still clueless, while two others have spotted the spell-casting necromancer, but the lookout in the foreground is the only one who's noticed the dragon might be waking up, which we're again indicating by the eye lights in the skull. And by using the same color for the light in the dragon's eyes and the necromancer's magic, we could really underscore the idea that the necromancer is the one in control here.

Now, as I mentioned, the narrative I'm constructing for each of these compositions involves at least some of the characters being unaware of the threat posed by the dragon. This is another technique that I lean on to expand the narrative weight that can be carried by a single image, because when the viewer knows something that the characters don't, it creates a tension that anticipates the character's discovery of the information the viewer already has. And this allows me to indicate character in its intentions and expectations prior to a watershed moment while simultaneously enabling the viewer to infer how the characters' emotions will change in response to this moment of recognition. So, essentially what we're talking about is a plot reversal captured within a single image. So in this next composition - and this is the last one, I promise - I distilled that idea down even further. I got rid of the treasure hunter subplot and replaced it with a direct confrontation between the necromancer and our player characters... who seemingly have him cornered. But the necromancer is smiling deviously in spite of his predicament because he knows he's about to get some major backup in the form of the fog-shrouded dragon he's summoning to protect him. Like the other comps, this one includes one of the player characters depicted in the exact moment that they realize the dragon is awake. In this case, it's the figure furthest to the right.

So this is basically a pictorial form of anagnorisis, which refers to the moment in a story when a character makes a critical discovery that changes the stakes emotionally. Including the story beat does two things. First, it adds an immediacy to the danger, because clearly the guy on the right is about to alert his friends and everything is going to change. And, second, it also makes it clear that the dragon is indeed a threat. Otherwise, it's possible that all these guys could just be ignoring the dragon for some reason, thereby undercutting his narrative significance, or perhaps the viewer could misinterpret the dragon is actually being allied with the player characters. The character experiencing anagnorisis tells us that his reaction is the same one that everybody else would be having if only they saw what he was seeing. It's a useful tool for clarifying character intent in complex compositions and expanding the dramatic tension of a scene. In any case, my art director liked this image, but he wanted it to have more story beats and to make it seem like our party of adventurers already had their hands full with other threats before the dragon even enters the ring. So, basically, he wanted to raise the stakes. So, I added in some reanimated human skeletons for our adventures to fight off, which also allowed me to promote the foreground figure on the left to our hero character still poised to confront the necromancer directly while his buddies take out the skeletons behind him. And then, finally, my art director felt that we needed more indications of the location setting for this scene, so I added in some architectural pillars to imply that we were in a massive subterranean ruin. So, with four potential options at this point I did a color toning pass on each comp, because while a composition may look great as a line drawing, when you introduce value and color, the readability and focal points of the image can change greatly, so it's important to block all this stuff in early-on so that I can catch any major problems before we invest too much time into a flawed composition. At this stage, the color tone mostly breaks down into warm and cool color influences meant to reinforce narrative cues like the relationship between the dragon and the necromancer as indicated by the same cool glow to their magic effects. Nevertheless, everything at this stage is still just placeholder and I typically end up redoing the line drawing, the grayscale values, and the colors once a specific composition has been chosen.

But once I complete the rough toning pass, my art director submits the options to the executive stakeholders, and then the other executives reply with questions, comments, and concerns about various sketches, and their consensus opinion results in one of three possible verdicts. So, first, they may reject all of the comps outright, in which case I needed to determine what it is about those options that I provided that's not working for them so I can do a brand new round of sketches that will hopefully be more successful. And this iterative, back-and-forth process can produce as many as three rounds of unique comps before one sketch is finally approved. And approval is the second possible outcome. But, more often than not, they deliver a third verdict which is a provisional approval, and basically this means that they like one of the comps, but they want me to make some changes to it before I start full-on production. So for the Dragon Bones key art we're looking at today, they did end up greenlighting the last of the four compositions I developed, although that's not always the case. So even though we had the greenlight to go ahead with production on this composition, we still wanted to refine some image elements a bit before I dove into detailed rendering work. One thing I wanted to do was introduce a warm light into the scene to contrast the cool shafts of moonlight coming in from above. And so, in order to justify that light source, I changed the pose of the figure on the far right and put a torch in his hand. And even though it isn't represented here yet, we also decided to add flames scattered across the ground between the necromancer and the front warrior which would add a warm fill light to pop that front warrior off the cool background behind him and also create a symbolic separation between the two characters. I also tried out a couple different poses for the necromancer and we ultimately settled on this one, because by dropping his left arm like this, we create some additional space in the image for the dragon's head - which you want to establish as a focal point - and it also opens the space between the necromancer and the front figure here so we can imagine that their eyes are locked on one another, reinforcing the story beat of the confrontation between them. Now, you might have noticed that throughout this entire sketching process the figures I've been drawing are all semi-naked bald dudes. Has anybody noticed that? That's because, in the early stages, it's simpler to sketch a generic human form to indicate the position of shapes and volumes rather than allow myself to get caught up in trying to capture specific details. And it's also because, even this early on in the painting, I'm using photo reference of myself to inform the poses for the characters. And you can see some of those references here. And there are a couple of benefits to using myself as a reference model, and the most important of these is that I have some idea of the mood that I want to convey in the pose, so I can kind of become that character myself without having to direct a model who has no idea how their pose fits into the image of the story of the illustration. But, I think the main point here is that having adequate reference is absolutely essential to producing high quality key art - and I would argue any art. And I lean my reference heavily throughout the entire painting process. Any day that I find I'm really struggling with what I'm doing, I realize it's because I don't have good reference, and then I have to go search for it or create it myself, and then things get moving again.

So, although all the figures look vaguely like me at the beginning, we obviously want to represent a more diverse cast of characters in the final painting. There are nine playable character races and two genders in ESO, and at this stage we decide on race and gender assignments for each figure in the painting. Of course, some of this is defined by the narrative. Like, if I'm depicting specific NPCs or if the scene is intended to take place in an area of tamriel where a specific ethnicity is dominant in the population there. We also want to represent at least one of the beast races, since they're a signature part of the Elder Scrolls series. And, in the interest of representation, when the narrative allows for it, I also try to represent a diversity of races from the real world as well as a mix of male and female characters. More generally, I just want to say that I believe that, as artists, we have a responsibility to be sensitive to cultural representation in our work and be responsive to prevailing cultural climates, which in turn our work helps to influence. On the one hand, it's so weird to think that my paintings of a fantasy medieval society have any relevance to contemporary culture at all, but the truth is, whether I intend it or not, every painting I produce makes a statement about how I personally view the world around me. And each painting also contributes to a cultivation of imagery that collectively shapes the way our players view the world too. And I'll give you just one example of how this has been relevant in my own illustrations. So, the second piece that I worked on after starting at ZOS was for ESO's Thieves Guild update. The update takes place in the province of Hammerfell, which is the home of a primarily dark-skinned people called the Redguards, and it was therefore natural that characters with darker skin tones would feature prominently in our key art. But the composition that developed during the sketching process depicted a lone member of the Thieves Guild evading city guards after pulling off a heist. I was working on this illustration in the months following the Freddie Gray protests in Baltimore and it occurred to me that it might not be the right cultural moment to showcase a black man being chased by law enforcement. On the other hand, however, populating the image with white characters in an in an in-game location known to have a light-skinned minority seemed inescapably like marginalization. So I voiced this concern and I even sketched out some alternative racial treatments for the primary character. But, initially, the executive team was focused on establishing continuity with the main character that had been depicted in an earlier illustration for the same update. So I continued with production with the character cast as a Redguard. And the image actually got fairly far into production before the decision was finally made to recast our hero thief as a Khajiit instead, which is the version that was eventually released to the public. Now, if we'd gone ahead with the original Redguard character, we may never have been called on it. The connection might not have resonated in the way that I feared. But the point is that with every decision you make in a painting it has the potential to make a specific statement to your audience, whether you intend that statement or not. So the most responsible approach is to be deliberate and circumspect with every choice you make in your art in order to avoid inadvertently making statements that you or your company might not intend.

Anyway, getting back to the Dragon Bones key art... Once the race and gender selections are chosen and approved, I start the linework stage of production where I go through each figure and modify the anatomy to match the body type of the race and gender intended for each character. I also draw clothing, armor, and gear on top of the anatomy drawings, and I tend to design the gear using inspiration from in-game armor sets which I combine with influences from photo references to inject more personality into the character's outfits. And, with the line work stage complete, it's finally time to start painting. So, as I mentioned before, most of the creative work for the key art is front-loaded. And, honestly, by the time I'm done with the drawing stage, most of the important creative decisions have already been decided on. My topic for today is not intended to focus on technique, but in the interest of being thorough and to give the session a satisfying conclusion, I'd just like to quickly walk through the remaining stages of my process so we can see how this painting finished up. Now, after the line drawing, my next step is to establish diffuse color values for the characters. So I use the lasso tool in photoshop to create selections of each material represented on each character. And then, on individual layers, I fill each selection with a unique color. So, for example, the skin will be on one layer, the leather armor will be on another layer, the metal trim will be on another, and so on. And the real value of this step is that it allows me to Ctrl-click the layer thumbnail of each material to restore the corresponding selection whenever I want, so I can use them to define masks for adjustment layers later on. Moving on from there, I render the lighting information in the scene in grayscale before introducing the color bias of each light source. And then, finally combining all of this with the diffuse material colors I established earlier, [sic] to achieve a rough approximation of the colored and toned lighting environment for the scene. Now, everything I've done to this point has really just served to establish an under-painting that I can use to start painting on top of. But, now, I can finally start the direct painting stage which makes up the bulk of the rest of the process. So, this is where I finally lose the drawing. I reduce the opacity of the line work layer and then begin painting in new layers on top of that. My process tends to involve a lot of adjustment layers which I prefer to keep live so that I can revise them later on, if necessary. And the downside to this approach is that it eats up a ton of memory. So, eventually, I end up using breakout PSDs to work on each character individually in separate documents. I focus on the characters one-by-one, treating each as its own mini-painting, before reincorporating them into the larger composition. And, in this case, the dragon counts as a character too. So, once I'm done with the characters, I focus on the surrounding environment, building in some features from scratch and also incorporating some in-game screenshots in the background that get heavily painted over to save on painting time. A lot of surface texture is generated through a combination of noisy brushes and photo texture overlays, and then in the later stages of the painting, I also do a final dirt-and-grime pass to help things feel convincingly aged and lived in. I add in cast shadows for the characters and special effects before doing a final color grading pass, which is often one of the most important steps of the entire process. This is where I finally establish a cohesive atmosphere for the illustration and rebalance the lighting and colors to increase readability and emphasize compositional elements in a way that better supports the narrative. And that is basically it. Aside from cleaning up the layer groups to help the marketing/web guys more easily navigate the file, I pass it on for publication, and then the whole process starts all over again for the next one.

So that is the end of what I had prepared for today. So, I just want to thank everybody for coming. I hope you found it worthwhile and I will be glad to take any questions that anybody might have.


First of all, I have no experience as an artist, but...

That's so much better.

This is incredibly insightful. I had no idea...

See, if you're not an artist, you can't tell me that I'm wrong about everything I just said, so...

Well, yeah, it's the composition. Anyway, sorry. I've noticed two things there's actually questions, hope this doesn't come off the wrong way, but in the initial sketches that you were giving, you have these concepts of like the Blackball and the Mandate, but you still didn't include characters in those initial sketches. Is that just something from your own point of passion that you're trying to be like "Hey, look, this is good. I need to see more of this." so that you can get it or is that something that you've now scrapped in your design process?

Yeah, that's a good question. I'm glad you asked it because it's something that I knew I had left out of this presentation. His question is "Why do I keep offering up sketches that don't include player characters from the beginning if it just creates problems for me later down the road?" And part of it is because we're doing this for three and a half years and over that length of time we've had different brand managers come in and out of the project on the Bethesda side. And so, they have different preferences. And so the concept of what defines a player character and how important that is has shifted over time. The other thing, too, is that - you know I mentioned three of those challenges, but there are a lot more that I you know weren't as important to talk about. One of them is that there are oftentimes specific angles on the story content that they're trying to push, and a lot of the time the two things contradict each other. So you can't tell those stories and also have player characters there that are doing something important. And so my initial solution is always to say "Look, we can do it this way." But there's this hard-and-fast rule that they typically come back to. So that's why.

So, in some ways, it's like a springboard for them to give you alternate feedback. Is that..?

Yeah. I mean, that's fair, yeah. I think so.

I know that I'm not supposed to ask follow-up questions, but there's only swords! Like how come nobody has anything but swords? I'm like looking at all this being like "Is that a hammer? No, it's a sword."

Well, so, a lot of what I sh - I think there are... Repeating the question, he says "There only swords." and "Why aren't we using other weapons?" I think there are examples of other weapons in some of the key art. The specific one that I broke down, it just worked out that way. I don't know good answer.

Yeah, I gotcha.

In a lot of the other sketches, though, I'll typically just use a sword as a stand-in weapon and then typically I will end up modifying the pose once it gets approved to a more specific weapon, because it doesn't make sense to bother with that kind of detail if the sketch is going to be thrown away anyway, so....

Well, thank you so much and great talk, thanks.

Yeah, thank you.


Hi. Great talk, great work, great project.

Thank you.

My question was in regard to getting started with your composition. I always find that the most difficult part of composition is getting started because when you get started it's not a composition yet.

Right, right.

It's a subject or it's a frame. My question was going to be "Do you start with a subject or do you start with a frame?" and it looked like, in the breakdown of those really simple sketches of the dragon, you were really focusing on the subject and not quite yet considering the frame because there was no outline on those sketches.

That is true. So yeah, I think the way I just work naturally, I will... "Do I start with a subject or a frame when I'm first beginning the compositions? Am I worried about how all those elements will fit inside a particular crop format?" The answer is no. Typically, just the way I work, I will start by drawing just a character on a massive PSD that doesn't have any bounds to it, really, and see where it takes me. And one of the reasons for that, in this case, is because the key art that we make - and I left this out of the talk, but there are a lot of actual format and crop specifications that I have to hit. And what that means is that I have to build a lot of margin area into every illustration that we do. So, oftentimes I end up painting this, you know, huge scene of a world that we only crop in on a small window of. That's actually - the illustration I'm working on right now is like that. Like, I've got so much that we're gonna have to crop in later and stuff outside of that will be wasted work, but that's just kind of the way it goes.

And I wanted to say as part of that: there is a sort of paralysis in the early phases of doing a concept. I'm not sure if you ever feel like you might be getting ahead of those requirements or desirements and whether or not you feel like "Whoa, I can't"... like there's a pressure to demonstrate that you're making something or that you can make something.

Uh-huh.

But also, you know that there's stuff out there that's not being expressed yet. You don't want to start getting ahead of that. So, like, how do you - is there a way in your process to manage that?

Well, I think I could take your question a couple different ways. But what I would probably say - I would rephrase your question as "Am I ever concerned that the people who see my artwork internally will make a judgment on the quality of the composition that I'm developing before I really get to a point where it's fully been expressed?" And then the answer to that is certainly "yes", but it becomes less-so over time because there is like a level of trust that gets developed and they just - they've seen enough examples, they know how it's more likely to turn out, and so they gave me the benefit of the doubt, so...


I don't know who was first, you do you want to go over here?

They always make these mics for tall people. So, my question... it is a technical question, so sorry. But I noticed that you had some meshes in some of your sketches and I was wondering if your 3d modeler gave that to you already rigged and you just put it in the pose you wanted or if they posed it for you or if you did it yourself.

Yeah...

And I also wanted to know what - you don't have to give me actual time/hours worked because I hate it when people ask me that question - but I want to know what your timeline in the whole production phase was.

Well, I would say overall what way typically gets slated that the entire production is supposed to last about eight weeks. And I do overshoot that sometimes, typically because there will be problems developed or the art director or someone else will decide that something's not working and then I need to go back and kind of, you know, retool the illustration and that adds time. So, sometimes it gets pushed. But eight weeks is what we shoot for in production schedule. As to your other question, which was "Where do the 3d meshes that you kind of see hints of in some of my sketches come from?" The answer there is: it definitely varies. So I started my career as a 3d production artist, so I certainly have the ability to build those myself. And I'll typically do that for environments or, you know, to build like the structure. Sometimes I'll do like a really rough gray box scene. Other times I'll go into ZBrush and turn on DynaMesh and sculpt like a face or some kind of object that I want to get from a different angle. And then, finally, in the case of what we looked at today, that dragon - I think is probably what you're talking about - there was a really early like uh proxy for the dragon, and that I did not make. But that was - it already existed and I just took the Maya file - or it was a Max file, but I work in Maya - I took the file and I cannot rig and it was not rigged itself, so I just basically used pivots and put it in the position I wanted and that is still faster than just trying to like draw anatomy that doesn't really exist anyway, so, you know, you don't want to go too far into debating with yourself whether or not it looks right. So, yeah.

Awesome, thank you so much.

Yeah, you're welcome.


And then over here?

Hi, Lucas. Thank you for the talk. It was really informative and it was really awesome, like your paintings are really awesome, I think.

Thank you.

So, just one small question is: When you're doing these key paintings, how important is it for you to do all the characters exactly as they appear in the concept part or exactly how they appear in-game? Do you have be so that they are like... faithful representations of those or are you able to also change these concepts to suit the needs of a certain painting?

Right. So, the question is: How much flexibility do I have in changing the appearance of the characters or other things for the key art even if they don't match the concept art or the final in-game models? The answer is: I have a lot of flexibility. And it didn't necessarily start out that way, but I think that my art director kind of has - he understands and firmly believes that the illustration is meant to be an aspirational depiction of what the game is as opposed to a realistic one. So I'm allowed to change things, make things look, hopefully, better and higher fidelity than they would in-game or in the concept. And also, I'm given a lot of flexibility with the costuming. So, oftentimes, I will use armor pieces from the game, but the thing is - about the way that we do armor sets in ESO - it's a very compartmentalized modular system. And so everything kind of has to look similar, it has to be very form-fitting to the character. We can't have fabric hanging off and things like that. And so, one of the things that I like to do is just use pieces of that, but build kind of something that looks more like a home-built kit, you know? Things that the character may have found and decided to incorporate into their outfit. And yeah, adding different kinds of wear and fabric hanging and things like that. So yeah, I definitely get flexibility when it comes to that. And rarely does it get pushed back - do I get pushed backwards like "No, you gotta make that look the same.", so.

Okay, thank you.

Yeah.


And, if that's it, thank you guys so much.