Interview with Kevin Roger, animation director on Sifu

As we all know, the development of a video game requires the collaboration of many different trades to bring to life the universes that fascinate us. From game designers to animators and scriptwriters, everyone contributes to the creation of a coherent whole. Of course, depending on the production, the importance of certain trades increases tenfold. Sloclap‘s work is a perfect example: 3D animation plays a key role in the experience offered by their games. With this in mind, today we welcome Kevin Roger, animation director on Sifu, for an interview focusing on his job and his role at Sloclap.

First of all, could you introduce yourself? What is your professional background and what led you to 3D animation?

My name is Kevin Roger, and I was born in the early 90s. I currently live in Paris, but my path began in the south of France, in Perpignan. To explain how I ended up in 3D animation, I need to share a bit of my story. When I was younger, I felt a need to escape, which was fulfilled by old family films and then by video games. It was impressive to see how these games transported me to another universe, and realizing that people were creating these worlds from scratch really made an impression on me. It made me want to do the same thing, to become someone who creates worlds for others to escape to. After following some advice that, in hindsight, wasn’t ideal, I ended up in IT, which didn’t really suit me. So I decided to change direction and enroll in an art school specializing in 3D filmmaking, even though I’d never really touched a pencil before, as I didn’t grow up in an artistic environment.

Following my training in 3D filmmaking, I was certain of one thing: I wanted to go into video games. To my great surprise, it was difficult to find my place in this sector, which I didn’t feel was that far removed from the world of 3D film. After about six months of searching, the Sloclap studio placed its trust in me and I’ve been with them ever since; it’s been eight years now. Today, they give me enormous confidence and allow me to express myself fully through my work. It’s also a studio to which I’m particularly attached, having played a part in its evolution. When I joined eight years ago, the studio had just been founded and there were just fifteen of us and 2 animators, including myself. Today, there are more than a hundred of us, including around fifteen animators.

So it was in 2016 that I started working on Absolver as a 3D animator, then on Sifu, where I had the opportunity to progress to the position of animation director.

We’ve noticed that you work in parallel at another studio called Arkada. How do you manage to balance these two activities?

Haha! Tricky question. Finding a balance between my work at Arkada and Sloclap was, in all honesty, a real challenge, especially in the early days. When I decided to help my partner with Arkada, it coincided with the launch of Absolver and both projects required a lot of attention. This undeniably had an impact on my health and social relationships, and was a period of intense learning to find a balance between professional and personal life, something I think many developers struggle with. Today, things with Arkada are much better. The project is up and running, and my involvement is less frequent as the team is more established. I’m only involved at critical moments, like when projects are launched, which means I can devote myself fully to Sloclap, where the ambition keeps growing and is more than enough to challenge me on my work/life balance, especially with projects like Sifu.

What are the differences between 3D and 2D animation?

The first major difference between 2D and 3D animation is rooted in their history. 2D appeared at a time when computers were just taking their first steps, requiring pencils, paper and a troupe of illustrators with clearly established functions. Production methods, calibrated to keep costs down, created a unique style in this art form. Then, technology evolved, applying these same methods in 3D and reducing the need for large numbers of animators and illustrators thanks to computer support. Today, distinguishing 2D from 3D can sometimes be subtle, especially as digital technology helps in both styles. For example, using a graphics tablet is an interesting in-between between traditional 2D and digital support for animators. 2D sometimes exploits the benefits of 3D for more advanced projects, while 3D is largely inspired by 2D. Sometimes, 3D film and video productions even seek to imitate the 2D rendering specific to the drawing technique, while retaining the flexibility of computers and 3D. It’s funny how cutting-edge technologies are developed with this objective in mind – it’s like a snake biting its own tail. Although we could discuss animation for hours, it’s vital to stress that, whatever the method or technique – 2D, 3D, stop-motion, etc. – animation is still animation. – animation is still animation. The fundamental principles remain fairly similar, even if the processes and workflows vary according to animator and technique. All in all, it’s a big, beautiful family that’s constantly evolving!

What’s the role of a 3D animator on a game like Sifu (and by extension Absolver)? What are their missions?

Even if Absolver and Sifu don’t look quite the same, they share a common core: their 3D animation. What’s more, both games share a common universe: the world of fighting games. For both titles, the 3D animation bridges the gap between game design and fantasy, in this case martial arts. The animators, with a host of inspirations at hand, observe, appropriate and adjust the movements to fit the often highly specialized and demanding game design. For Absolver, which is more player vs. player, the opponent’s intentions and movements need to be extremely clear, so that the players can react, defend themselves and return the favor. The same principles apply when our opponent is an AI, although players are more willing to accept surprises and, for some, enjoy learning sometimes vicious patterns. In a versus game, we need to avoid frustrating players and ensure consistency within the combat experience.

With Sifu, the team really sought to achieve a high level of authenticity, making sure to respect as much as possible the martial arts that inspired each of the characters’ movements. As a result, the interpretation of the movement had to be even more exacting.

To sum up, the 3D animator has to convey the constraints of the game designer through movements, while adhering to an artistic animation direction. Faced with a wealth of super-precise technical information, such as the time to impact, the origin of the blow, the limb used… , the animator has to think (a real headache) about how to subtly camouflage these requirements in the movements, while drawing on existing martial arts to keep things natural and fluid. Everything else apart from the combat is fairly classic: producing all kinds of animations in line with an artistic direction, and thus dressing up the characterization of the characters.

What techniques are used on Sifu to animate the characters?

All of Sifu‘s gameplay has been meticulously animated by hand – “Keyframing”, “Handkey”, whatever you want to call it. The animators define the characters’ poses frame by frame in 3D software, based on reference images. This is a testament to the expertise of Sloclap’s animation team, which I believe gives our projects their own unique animation identity.

Animation is important to Sloclap. We devote a lot of time to it, and we often get together with other animators to exchange ideas, give each other feedback and talk about animation. Although the studio’s approach is based on handkey on Maya (3D animation software), we also use motion capture at times, but only in a minimal and targeted way. This technique has been used as a sketch or initial model for cinematics or animatics, helping us to quickly confirm or reject certain creative intentions.

In addition, we used motion capture for the “finishers” we call “takedowns”. It’s crucial to note that these animations were substantially reworked by hand, to the tune of around 80%, in order to harmonize the artistic direction of the animations and maintain consistency with the rest of the gameplay, which is entirely hand-animated. In short, motion capture, while useful for sketching complex scenes, has been methodically remodeled to optimize its impact and cinematic appeal.

From your point of view as a 3D animator, what have been the contributions of highly accessible game engines offering a wide range of tools, such as Unreal? Have they profoundly changed the way you work with 3D?

Game engines, such as Unreal Engine, are now easily accessible and packed with great tools for animators. For example, Unreal offers the possibility of viewing Maya animations live in the game engine, thanks to the “livelink”. There are also features that improve working conditions, such as “control-rig”, which allows animators to animate directly in the game editor, without having to use third-party software such as Maya or Blender. Of course, these tools have their limitations, but they’re evolving very fast, and it’s really great to be able to do all this directly in the engine.

What’s more, the “sequencer”, which acts like a non-linear video editor in the manner of an editing software, lets you play around with a whole host of things, including making cinematics or small videos with camera movements etc. very easily. Unreal even lets you play with state machines to animate a game character with zero programming knowledge, and see the results very quickly. This free access leaves more room for creativity, to perfect animations and projects, try out new ideas, improve and make yourself known to studios.

Of course, the web is also overflowing with resources, which is a real training ground for those learning 3D animation and the use of game engines on their own. Finally, although Unreal is often highlighted, let’s not forget other excellent tools for 3D in general, such as Blender, which is increasingly accessible and robust, and offers a great starting point for those new to 3D animation. However, it’s essential to note that, while these tools drastically facilitate and accelerate the learning process, they don’t replace the complexity of the medium or the need for intense practice. 3D animation remains a demanding field, requiring continual dedication and exploration to fully master the art and technique.

Did you develop specific in-house tools to help you with the development of Absolver and then Sifu? If so, how did they evolve between these two games?

Of course, customization and the development of specific tools are essential in our industry, whether in animation or any other craft. Each studio, for each project, must adapt existing tools, modify them and create new ones to meet specific needs. Although I can’t provide an exhaustive list of all the tool adaptations and creations we’ve carried out, the example of our combo editor is quite evocative. This tool allows us to iterate quickly and easily by creating movement sequences for our characters directly in the game editor in the form of nodes connected to each other. With regard to animation specifically, we began developing our own tools to facilitate character creation and manipulation by our animators, in order to optimize productivity, avoid repetitive tasks and minimize errors. These tools evolved throughout production. The phase when these tools undergo the most development is pre-production, and with each new project, we draw on the experience gained to improve and adjust them for future projects.

Sifu’s concept is original, to say the least: the character ages as he dies. Does this translate into concrete gameplay in terms of character animation?

Ouch! The knife in the wound! Haha! Unfortunately, in the current version of Sifu, the aging of the character has no influence on the animations.

Although the idea was considered during pre-production and initially excited us, imagining movements that evolve in precision and efficiency with the character’s age and experience, it never materialized. The reality of video game production for a studio of our stature, particularly in terms of costs, made this ambition unattainable, especially when we set it against the added value and perceptibility of this element for players.

Cases of major ambitions being either scaled back or cancelled are a regular occurrence, and sometimes represent several weeks of work thrown away when we fail to correctly anticipate all the challenges induced by the addition of a mechanic or concept. It’s a titanic task to predict this kind of thing several years in advance. Even with experience and time, it’s always complicated to anticipate, because every production is different. There are always unforeseen circumstances, and almost every day you have to come up with solutions, make trade-offs and take risks within a project.

Ultimately, a game production is very organic, which is what makes it fun in my opinion. You might think of extracting a pattern from one experiment and applying it to the next. This is possible in theory, but not in practice. All the same, we learn to be a little more reactive, and our judgement sharpens as to the choices to be made to preserve the core of the experience. Of course, when these choices are passed down the chain, often to junior developers (the beginners) who have put a lot of energy and love into their work and become attached to it, this generates a great deal of frustration. So, tact and appropriate human management are needed to prepare teams for this kind of event, and to teach them not to identify with their work more than they should, in order to avoid too much frustration and loss of motivation, and so on.

On a personal note, my first cut was a weapon in Absolver, which I was particularly fond of and which took me over 3 months to prototype. Worst of all, it was I who confirmed the cancellation, bearing in mind the time and effort involved.

Because of its concept, the visual reading of animations by the player must be particularly precise. How did you achieve this without using visual indicators (or at least as little as possible)? Did you have to make concessions between legibility and the realistic/plausible aspect of the animations?

Animation, especially when it comes to martial movement, is intrinsically subtle and complex. Each animator can bring his or her own nuance, enriched by experience.

Let’s take an offensive attacking movement: it can be segmented into three key parts, namely build-up, force generation leading up to impact, and release, which illustrates the consumption of the energy committed to the movement. Each of these phases is vital to ensure that the player can visually “read” the animations:

• build-up, for example, signals the intention of the movement, giving crucial indications of the direction and limb used, all before the movement is initiated,
• impact, on the other hand, can be shaped as a stylistic vector for animation – vivid and “snappy” if the anticipation is clear and explicit,
• the release, on the other hand, provides the player with a window into the power and commitment of the move, informing them of the impact of the move on the character, and signalling when the player will be able to take control again.

Avoiding player frustration is crucial, and this is achieved by ensuring that animations faithfully reflect the intentions and possible actions dictated by the game design. Careful selection of reference movements, followed by in-depth work on elements such as the build-up, playing with silhouette and timings to highlight the limb used or the origin of the blow, is paramount.

On Absolver, there was more freedom in the interpretation of reference movements, whereas with Sifu, authenticity was more in demand. In real life, you don’t necessarily want to “telegraph” your moves and reveal your intentions to your opponent. In the video game context, this is a necessity, but it has to be done subtly so that the player can understand and enjoy every move. To this end, we consulted a martial arts expert to help us choose movements in line with our design intentions, minimizing the need for interpretation in the animation. Secondly, there’s no secret: iteration is fundamental to refining the balance between authenticity, animation style and design constraints: each of the attack movements can be reworked dozens of times at any point in production to achieve this balance.

In Sifu, each boss is unique and represents a real challenge for the player. From your point of view, how much does the animation of these characters add to the uniqueness of the characters in the game?

In Sifu, each boss illustrates a tribute to a distinct martial style, echoing our main character’s practice of “PakMei”. Fajar practices Pensak Silat, Sean practises Bajishuan, JinFeng masters Long Fist kung-fu, Kuroki skilfully mixes Okinawan Kobudo and Kenjutsu styles, and finally, Yang, the ultimate boss, uses the same martial art as our hero.

These artistic decisions are partly dictated by their backstories and the weapons they wield. In our quest for authenticity, we have chosen emblematic movements for each martial art, and the staging, combined with the dialogue during cutscenes and fights, completes the identity and impact of each character within the game.

Today, in many fields, “fans” have an influence on creation, whether by praising it or by the toxicity of their behavior. Absolver unfortunately received a mixed reception from gamers. Was this a pivotal factor in the move towards a single-player game with Sifu?

It’s interesting because for us, the reception of Absolver was actually quite encouraging and reassuring enough to allow us to continue along this path.

Considering the context of an emerging studio with a relatively small and junior team, coupled with the complex yet ambitious project that was Absolver, we were delighted to see the game find its audience and see a community build up around it. As a result, the game’s reception didn’t really influence our future choices. Our desire to continue exploring martial arts came from a passion, and we wanted to do it in a slightly more cinematic and visceral way, in homage to classic martial arts films. We wanted to immerse players in this universe. Multiplayer seemed superfluous for this type of experience and, frankly, by sparing ourselves this technical complexity, we felt more comfortable trying to offer a deeper, more accomplished experience.

The combat system, based on dodging, timing and special moves, reminded us a lot of Clover Studio’s God Hand. Was this Beat’em Up a source of inspiration for you?

Great culture and observation! God Hand is indeed a title cherished by Sifu‘s creative director, who was also game director on Absolver. This game has been sitting in our reference drawer since Absolver and has constantly been one of our main inspirations. Not necessarily in terms of the animation or the tone of the game, but more in relation to the gameplay mechanics, as you mentioned. God Hand has influenced us in certain ways of approaching combat and interaction in the game, with an emphasis on dodging, timing, and special moves for sure.

To facilitate your work on the animations, did you need to watch any action films? Did films like John Wick, or the Bruce Lee classics provide a basis for your work?

When it comes to animation, you need references. These can come from everyday observations, news reports, fiction, YouTube videos and so on. It’s a fundamental aspect of the animator’s craft to analyze, interpret and apply these observations to our art.

Of course, we watched a wide range of films to enrich ourselves:

• Drunken Master and Police Story with Jackie Chan,
• The Way of the Dragon and Game of Death with Bruce Lee,
• Once Upon a Time in China and Kiss of the Dragon with Jet Li,
• Ip Man, of Iko Uwais and The Raid with Donnie Yen,
• Ong Bak, Kill Bill, The 36th Chamber of Shaolin with Tony Jaa,
• more recent action films such as John Wick.

These works were all sources of inspiration for Sifu. We have also drawn on videos of Wushu martial competitions and online martial arts courses. The elements we look for in these references vary: we may seek to understand body mechanics, perceive the subtleties specific to each practice in a search for authenticity, understand the applications of certain techniques, appreciate the aesthetics of choreography, feel the viscerality of combat, perceive the cinematographic aspect and enhancement of movement, and more.

Once again, we’d like to thank Kevin for his availability and trust.