When an Early Access Character Design Broke My Trust

When an Early Access Character Design Made Me Lose Confidence in a Game

Before reading further, look through the screenshots above and form your own opinion. What feature do your eyes notice first?

This article isn't an attempt to prove that a character model is objectively bad.

In fact, the reason I'm writing it is because I couldn't prove that.

What I had instead was a strong reaction.

The first time I saw this character, something immediately felt off.

I wasn't measuring facial proportions. I wasn't drawing guide lines. I wasn't comparing screenshots.

I simply looked at the character and felt that something about the face was undermining the attractiveness the design was clearly aiming for.

More specifically, I couldn't stop looking at the mouth.

That reaction eventually led me into Discord discussions with the developers, comparisons between pre-fix and post-fix versions, and a surprisingly deep rabbit hole of facial analysis.

What I do know is that the experience changed how much confidence I have in the game's artistic direction.

Here's the context for the slides:

• Slide 1: A screenshot of Mizuki, the character found near the broken-down car in the Early Access version of the game. (To find her, head to the car dealer to rent a vehicle, which triggers the side quest to locate the car.)
• Slide 2: A side-by-side comparison of Mizuki before and after the developers' updates. Personally, I didn't see any noticeable difference between the two versions, so I asked if they had accidentally uploaded the wrong "after" picture.
• Slide 3: Their actual response to my question.

The Problem: I Couldn't Stop Looking at the Mouth

Most character artists want players to notice the eyes first.

Eyes communicate personality. They draw attention. They establish emotional connection.

Yet whenever I looked at this character, my attention kept getting pulled somewhere else.

The mouth.

I wasn't consciously analyzing it at first. It simply dominated the face.

The more I looked, the more I felt that the smile occupied an unusually large amount of visual space relative to the rest of the facial structure.

The lips appeared broad and prominent. The smile stretched widely across a relatively delicate face. The visible teeth amplified the effect further.

Most importantly, the mouth became the focal point of the character whether I wanted it to or not.

That isn't necessarily a flaw. But it was the beginning of my discomfort.

In an adult romance game, player immersion hinges entirely on attraction and emotional connection, which we naturally seek through a character's eyes. When the visual hierarchy is completely hijacked by a tense, dominant grin, it triggers the 'uncanny valley' effect. Instead of feeling drawn in, the brain registers a biological red flag—the face looks less like an attractive partner and more like a rigid, robotic mask.

Trying to Find an Objective Explanation

After staring at screenshots for far longer than any sane person should, I started trying to understand what exactly was bothering me.

A few possibilities stood out.

• The mouth appears visually large relative to the rest of the face.
• The lipstick increases the apparent size of the smile.
• The smile exposes a significant amount of teeth.
• The expression lacks a true 'Duchenne marker'—meaning the eyes and upper cheeks remain wide, static, and completely unblynking while the mouth stretches. A real smile crinkles the eyes; a forced or uncanny smile leaves them hollow.
• The distance between the nose and mouth feels relatively short.
• The horizontal stretch: The lip corners are compressed and pulled straight backward rather than curving naturally upward, turning a flirtatious gesture into an aggressive, horizontal grimace.

None of these observations prove that the model is incorrect.

Human faces vary enormously. Stylized characters vary even more.

What they do provide is a possible explanation for why my brain kept flagging the face as unusual.

The issue may not be any single feature. It may be the combination.

The Shared Animation Caveat

Given the developers' point about shared animations (see Slide 3), there is a very high probability that this uncanny effect is unique to Mizuki. In an indie pipeline, a single "big smile" animation asset is often mapped across the entire character roster to save time. It is entirely possible that this animation looks perfectly natural on other characters in the game—either because their base facial geometry happens to tolerate the structural stretch better, or because their specific personalities simply don't require them to smile so broadly.

However, that is exactly the danger of a cut-corner pipeline. A shared asset might pass internal QA on 80% of the cast, but the moment it is forced onto a character with a slightly different jaw structure or lip width, the geometry completely breaks. Mizuki essentially became the canary in the coal mine for the engine's rigging limits.

The Realism Problem

The most interesting explanation I found had less to do with anatomy and more to do with artistic consistency.

Most of the character seems to be pursuing a semi-realistic aesthetic.

• Realistic lighting.
• Realistic skin shading.
• Realistic hair rendering.
• Relatively realistic facial proportions.

Then the smile arrives with a level of exaggeration that feels as though it belongs to a different artistic language.

Cartoons can get away with enormous smiles.

Animated films can get away with enormous smiles.

Highly stylized anime designs can get away with enormous smiles.

But those worlds establish stylization everywhere.

Here, my perception was that a mostly realistic face suddenly switched into a more cartoon-like expression.

The result wasn't charming. To me, it felt jarring.

This clash of styles is made significantly worse by the game's engine environment. When you see her out in the open world, flat daylight strips away any forgiving shadows. Without curated studio lighting to soften the edges, the base geometry's flaws are laid bare: an idealized, attractive body model completely at odds with a rigid, unexpressive facial mesh.

The Clown Comparison

This is the section where I expect some readers to completely disagree with me.

The smile reminded me of an exaggerated clown grin.

Not because the character looks like a clown. She obviously doesn't.

Rather, because clown makeup often works by visually expanding the apparent width and prominence of the mouth.

It takes one of the most emotionally important features of the face and amplifies it.

The combination of broad lips, visible teeth, and strong emphasis around the mouth triggered a similar reaction in me.

Instead of seeing an attractive face first, I saw the smile first.

And because the smile felt exaggerated relative to everything around it, it produced an uncanny effect.

Your mileage may vary. Mine certainly did.

The Discord Discussion

What surprised me most was not the lack of agreement by other players. People disagree about art all the time.

What surprised me was how little reaction there seemed to be.

I raised the issue because it felt obvious to me.

Not subtle. Not debatable. Obvious.

Yet the response was largely indifference.

The developers eventually produced a revision, but the before-and-after images looked so similar that I struggled to identify meaningful changes.

Eventually the design was declared final.

Case closed.

Except it wasn't closed for me.

Why This Became a Trust Issue

What ultimately bothered me wasn't the face itself.

Artists and players disagree all the time.

What bothered me was realizing that the thing I immediately perceived as a problem either wasn't visible to the developers or wasn't important to them.

The developers explained in their Discord reply that the game currently relies on shared animations, claiming the smile only looks wider due to the screenshot's lighting. But this defense completely misses the point.

The root issue isn't just a lighting fluke; it's a structural failure of asset sharing. When you force a generic, one-size-fits-all smile animation onto a unique face mesh without adjusting the underlying rigging, the geometry breaks. For a game whose core gameplay loop relies heavily on the physical appeal of its characters, cutting corners on facial rigging for a quest and potentially 'romanceable' character shows a fundamental gap in artistic priorities.

That's where the discussion stopped being about a mouth and started being about trust.

Early access development is built on trust.

Players aren't only buying the current version of a game. They're buying into a vision of what the game will become.

When a design decision immediately jars me, and the developers seem satisfied with it after only minor adjustments, I start wondering whether we share the same artistic priorities.

If we don't, that's not necessarily a criticism of them.

It may simply mean that the game is evolving toward an audience with different tastes than mine.

Even if we give the developers the absolute benefit of the doubt and assume they simply uploaded the wrong 'after' picture by mistake, the trust issue remains identical. The fact that no one internally caught such a glaringly identical 'before and after' comparison before publishing speaks to a severe lack of quality control. Furthermore, their immediate response wasn't to double-check their files—it was to defensively tell the player that they were the only one complaining and that their perception of the character was wrong. Whether it's a failure of asset management or a failure of basic anatomy, it points to a team that is fundamentally disconnected from the visual standards their genre requires.

As a player, that's still valuable information.

Is It Just Me? (And the Presentation Problem)

Of course, artistic appeal is inherently subjective. In the broader landscape of indie adult games, players often operate on a sliding scale of tolerance for technical jank. For many, idealized physical attributes are enough to satisfy the genre's baseline requirements, and the fact that the rest of the female cast uses a similar, safer baseline mesh means Mizuki is the only one who truly pushes the boundary into the uncanny. If you view her strictly from a distance, the mesh technically does its job.

But context is everything. The developers argued that Mizuki "does not always look at you with that big smile," implying the expression is fleeting. However, in a visual-novel-style dialogue system, characters pause and hold their expressions while the player reads the text. You aren't catching a split-second frame of animation in motion; you are forced to stare directly at a static, unblinking grimace for blocks of dialogue at a time.

When a game relies on close-up, narrative immersion to build attraction, forcing the player into a prolonged staring contest with a broken facial mesh isn't a minor subjective hitch—it’s an immersion breaker. Ultimately, when the core loop of an adult game relies on visual appeal and the response to basic feedback is defensive pushback rather than quality control, it leaves the player with only one real recourse: hitting the refund button and uninstalling.