How to Create Fun Stylized Halloween Scene in 3D
Sara Kołodziejak and Adrian Horodyski told us about their Halloween project, discussed their shared art direction, and shared some lighting tricks.
That project was also a collaboration, since we worked on separate characters within the same universe. During the Hades challenge, we created characters that relied almost entirely on hand-painted textures, so to make this project feel different, we wanted to build a small scene where lighting and atmosphere would play the main role, not just the textures.
In this article, we’d like to share our workflow for a project that we created together from start to finish, and show you how we worked on characters, lighting, and materials for this fun scene.
Maybe They Just Want Toilet Paper?
We did our first collaborative project in 2021. While working on it, we realized how well we function as a team. We share a similar stylized approach, are drawn to the same aesthetic, and our skill sets complement each other naturally. Since then, we’ve been creating a new Halloween project every year. During these projects, we always try to learn something extra, something outside our usual day-to-day character work. That’s why we keep expanding these Halloween pieces into full scenes, sometimes even adding small environmental animations.
In 2024, we even collaborated with our super-talented animator friend Pereiranimations on a little looped animation for a Have a Nice Death fan art.
We keep a dedicated space where we collect ideas and concepts throughout the year, exclusively for Halloween, and sometimes it takes months before we land on an idea that excites both of us. It’s important for us that each project is a different style so that we can explore and challenge ourselves.
For this year’s project, the starting point was an amazing illustration by Alattecrap. The moment we found it, there was no debate; we are doing it! That tense atmosphere, the emotions, and those adorable school kids characters themselves, we love everything about the original piece.
The division of tasks also came naturally: Sara claimed the two girls in the foreground, and Adrian took on the boys. When it came to the environment, the distribution was just as straightforward: Adrian worked on the background models, Sara handled the textures, and we collaborated closely on lighting and the overall polish of the scene.
The biggest challenge in collaborations like this is maintaining a consistent style, proportions, and interaction between all the characters. That’s why we needed to share progress frequently, constantly checking how our models worked together and making adjustments as needed.
Our main reference, of course, was the illustration itself. However, we also gathered additional references from other artists whose style we enjoy, mainly to establish the direction for the scene and the characters, so we could clearly define the style we wanted to achieve from the very beginning.
Sara: I also find it very useful to spend some extra time on a concept, analyzing the shapes, breaking them down, and drawing over them. Here, I had trouble understanding the facial features. Since Adrian’s character had a similar expression, I decided to give the front girl an open mouth too, even tho we weren't fully sure that it was the same as on the original piece.
Adrian: Before starting the sculpt, I always spend extra time closely analyzing the concept’s shape, language, and use of negative space. In this illustration, I noticed the artist heavily relies on angular forms, both in the clothing and in the negative space around the characters, and even in the angled zombie arms on the ground. Keeping this visual language in mind helped guide many of my decisions throughout the sculpting process.
One of the main challenges I faced with the male characters was determining the blonde boy's body position. It’s hard to understand what’s happening with his hands just from the illustration. Since we wanted the camera to rotate around the characters, their poses needed to work from every angle.
I explored several options, and the best solution was placing his hands against the wall. This allowed his arms to align naturally with his legs, maintaining the pyramid-shaped composition from the original illustration. Overall, I was very drawn to the energy of the concept, and my goal was to capture as much of that dynamic feeling as possible in our recreation.
Modeling
We always model our characters in ZBrush, then we move to Blender, where we work on the rest of the scene and renders. Here, just before moving to ZBrush, we used poseable mannequin models to establish the general placement of the characters, then we created a simple bathroom environment and the camera.
Setting up the mannequins as a sketch pass for the characters was extremely useful as it allowed us to work on our models independently while ensuring, from day one, that the proportions, positioning, and interactions between them would align correctly.
We both enjoy character work the most, especially in ZBrush, so that’s where the core of the process took place. We each started by blocking out our characters individually. Whenever one of us made a major change, we exchanged files, merged everything back into the shared scene, and continued iterating together.
The character sculpting process in ZBrush was fairly standard. For sculpting stylized folds on the clothing, an effective method involved masking the area with Lasso Mask, then pushing it in using the Inflate/Move brush, enhancing the crease with the Gio brush (the best brush ever, shout out to Gio Nakpil), and finally cleaning everything up quickly with ClayPolish.
Texturing & Scene Setup
We didn’t need these characters outside of this specific scene, so topology wasn’t something we needed to worry about. The characters use very simple texturing, just ZBrush polypaint, and to keep it even simpler, we reduced it to just base colors with just a little color detail on eyes and faces.
For our purposes, that was more than enough: we simply exported a decimated ZBrush sculpt, plugged the texture in as a Color Attribute, and adjusted the roughness, metallic, etc. We also added a simple, thin outline with a grease pencil.
For the environment, however, we did require UVs and proper topology so we could apply textures like graffiti on the walls.
The wall texture was a combination of a base color and an alpha texture on which we painted the graffiti, inspired by the details from the original illustration. The stickers were created as separate images with alpha and given a slight thickness to make them feel more integrated into the scene.
For the zombie hands, we first created a single base model and posed it using a simple rig. At a later stage, we did an additional sculpting pass and added more detail to refine the final look. We used two layers of different noise textures that were mixed.
For extra bumpiness, we used a stretched Vornoi texture. We also added slight subsurface scaterring. We decided to model extra slimy strands coming from hands, to add a bit of gross factor to it, and to make those hands feel more integrated into the scene.
Lighting, Rendering & Final Presentation
Our lighting setup is fairly complex, but that’s because many objects have their own individual lights that affect only a specific asset. For example, the rim light on the character’s hair wouldn’t be possible without also hitting the narrow walls behind it, and those reflections could easily disrupt the whole scene. That’s where Blender’s Light Linking feature turned out to be extremely useful, letting us include or exclude specific objects from being affected by certain lights.
Another lighting trick we used was adding a very thin volumetric fog across the entire scene, including the camera. It slightly softened the shadows and added a nice bit of light scattering overall.
We decided to render in Cycles because we really wanted that cinematic look, and the few reflective materials we made for the walls and floor looked much better in Cycles than in Eevee. The same goes for the subtle subsurface scattering we added to the zombie hands material.
Initially, we established the camera placement with fSpy software. By analyzing the vanishing lines on the reference image, fSpy accurately reconstructs the camera's parameters (focal length and orientation), allowing for direct export to Blender. That initial camera placement was very helpful for quickly achieving the correct perspective, even if later we changed it slightly. Overall, fSpy is super helpful in figuring out the right camera placement for even more complicated scenes.
Final Thoughts
The most important lesson this project taught us was the good old “trust the process”. Before anything starts looking appealing, there’s often a long period filled with very rough, very unflattering blockouts. Those are the moments when you feel tempted to drop the project entirely, because it’s hard to imagine a polished final result. But you have to take it one step at a time, comparing your progress to earlier stages rather than focusing solely on the end goal.
Other than that, we strongly recommend two Blender-oriented creators: Ucupumar, whose addon Ucupaint lets you texture in Blender almost like in Substance 3D Painter, which we used for some last-second texture adjustments, and SouthernShotty, whose YouTube tutorials on stylized materials and scenes are incredibly helpful.
In total, it took us around a month to complete this project. As always, it was a lot of fun to work together. Collaborative projects are great, especially when you can divide tasks in a way that lets each person focus on what they enjoy and do best. We both love working on characters the most, so in any project we tackle together, each of us must have our own character to develop. But when it comes to the remaining tasks, we always try to divide them evenly.
Thank you for reading, and if you are interested in more, check our socials. We are already gathering ideas for the Halloween 2026 project!
Sara: Instagram, ArtStation, and X/Twitter.
Adrian: Instagram, ArtStation, and X/Twitter.
