Discord Link Here :
In the first episode of “How We Made Firva as an In-Game Character,” we went over the major design ideas behind Firva and explained how we organized the full journey into five main groups (you can revisit them here):
1 - The Cel-Shaded Group
Our earliest complete designs, fully covered in Devlog #07.
2 - The Early Realistic Group Our first steps toward a more grounded visual style, analyzed in this Devlog.
3 - The Transition Group
A series of iterations exploring hybrid looks and new ideas.
4 - The Experimental Group Designs where we tested bold new directions and pushed boundaries.
5 - The Final Direction The point where we reached our full visual and narrative potential.
When we decided to change our entire art style, one of our biggest concerns was how to represent Firva, the main character in this new visual form. At that stage, we still hadn’t figured out how realistic we wanted the game to be, or even what type of realism would fit the world. So the character redesign wasn’t just a cosmetic update; it was a foundation-level problem we had to solve while everything else was still shifting.
Because of that uncertainty, it was imperative that we explored a wide range of possibilities early on. We knew that if we locked ourselves into a single direction too fast, we’d risk running into major issues later. (Spoiler alert: We still ran into a lot of issues, but exploring broadly at least helped us understand the set of problems we were dealing with.)
At the same time, we were also researching pre-made art assets to see if we could speed up development we did not have any artists on the team, so this was our only logical choice. Since we were looking into potentially integrating external resources, we had to make sure their artistic language matched the direction we were evolving toward. It wasn’t enough for an asset to look “good”—it needed to belong in Firva’s world, and align with the style we were aiming for, and vice versa, Firva herself must also belong to that world in a believable sense.
Another huge factor was our team situation. We had very little to No experience with full-scale character art production, and after our artist left, we were a small three-person team trying to figure out many of these things on our own. So even with all these limitations—uncertain style goals, the pressure to move fast, limited manpower, and limited character design experience—we were still completely committed to pushing forward.
The first practical step we took was switching to a different 3D character workflow entirely. Since we weren’t trained artists, we needed software that was easy to learn, fast to iterate in, and reliable enough to get us to a high-quality character without a long production pipeline. After looking around, we landed on Character Creator 4 by Reallusion.
Character Creator 4 focuses specifically on realistic character creation, and at the time, it was the most capable tool we found for producing high-quality human characters quickly. For a small team like ours, that mattered a lot: we needed something that would reduce friction, shorten iteration time, and help us reach a believable result without needing a full character-artist skillset from day one.
After a lot of tweaking and countless small adjustments, this new version of Firva finally came to life. (And sorry for the poor quality here—this is the only footage we still have from 2023–2024.)
(This version Is NOT in USE)
With this redesign came a wave of new ideas, most of them pulled directly from the updated story we were developing at the time. This Firva was noticeably older than the cel-shaded version—roughly 16 to 18 years old—and her entire vibe shifted with that. She was l meant to be edgy and sharp this time, still carried the red hair we considered a core part of her identity, but her backstory changed dramatically.
In this version, Firva had spent some time living alone in the jungle. That single narrative change affected almost everything about her design. Her clothing, for example, moved away from the “high-born” look we had in earlier concepts. The elegant, noble influence was stripped out and replaced with something more survival-driven: leather armor, rougher materials, and a silhouette closer to a rogue-style character—someone who’s adapted to the wilderness through necessity, not choice.
We also revisited an older gameplay idea here: giving Firva a smaller dagger-type weapon as her main tool in combat. We wanted to see how that choice would feel with her new proportions, animations, and overall style—whether it still fit her identity and if it could support the kind of fast, aggressive fighting we were imagining.
Below, you can see an early test of the new fight system for that time, along with this realistic Firva in action. It’s rough, but it captures the direction we were aiming for at the time and shows how these ideas started to connect in motion.
(This version Is NOT in USE)
(This version Is NOT in USE)
Then came the truly difficult part—the nightmare of many indie devs: the almighty hair physics. And honestly, this is where Firva’s “edgy” look started to fall apart for us.
The first hairstyle we chose just wasn’t working the way we planned. No matter what we tried—different meshes, different physics settings, different levels of detail—it kept failing in one of two ways. Either:
1 -
It looked ugly / unnatural in motion, breaking the vibe we were going for, or
2 -
It looked decent, but was way too expensive in terms of processing power and performance.
The hair by itself contained 1.5 Million polygons !!!!!!!!!!!!!!!!!!!!!!!!
And we weren’t exactly hair-tech experts. Maybe it was a lack of knowledge, maybe it was how the tools handled simulation at the time, or maybe it was simply the limits of our pipeline—but the result was the same: we couldn’t get the edgy hairstyle to behave and stay within our performance budget.
So we had to make a call. Instead of forcing a style that kept fighting both the engine and our skills, we toned it down. We simplified the hair direction and started leaning toward designs that were easier to simulate, easier to animate, and more stable in gameplay.
It was one of those moments where reality punched the concept in the face: what looks great in a still render can become a disaster once it has to run, move, collide, and survive real-time combat. Hair wasn’t just decoration anymore—it was a technical system that could derail production if we didn’t respect it.
This video is also the same fight test but with the final version of firva for The Early Realistic Group
What Comes Next
In Episode 3, we’ll dive into the Transition Group—the stage where Firva went through the most visible evolution. This is the part of the journey where you’ll meet the version of Firva we ended up using the most during development, along with the chain of iterations that led to her.
You’ll see how that “core” Firva design didn’t appear all at once, but slowly took shape across multiple versions—each one solving a problem, introducing a new idea, or pushing the character closer to the final direction.
So yeah… that’s where things get really interesting.
Stay tuned.
Join the Conversation
We’d love to hear your thoughts, questions, and feedback. Join our community on Discord →
or leave a comment below