Sumo Girl

Goals and Objectives

Complex conceptual exploration at the intersection of different styles and techniques is my favorite type of work.

As an example, I’ll describe the process of creating a female wrestler character, which I developed as part of a small series of personal projects depicting various types of strong, athletic women.

3. Software

• Pureref
• ZBrush
• Maya
• Advanced Skeleton
• Wrap 3D
• Marmoset
• Photoshop
• Lightroom
• Snap Art
• TVPaint
• DaVinci

References & Planning

People are the most common and interesting subjects in art. In this series, I aimed to create and showcase different types of characters in an engaging way, staying as expressive as possible, and relying on my previous works and a wide range of references.

One of these characters is a girl inspired by female sumo wrestlers and other women involved in weightlifting or other heavy athletics.

My character wasn’t meant to be a documentary-accurate portrayal of such athletes, so I allowed myself some creative freedom in the costume and hairstyle.

I also aimed for stylization and simplification in the anatomy and textures. The final renders were intended to look like a combination of 2D and 3D, referencing the styles of Arcane, Spider-Man, and more broadly, the drawing techniques typical of 2D animation.

Before starting any work, I strive to envision the final outcome as clearly as possible.

While much can be adjusted later, it’s always beneficial to have a vision of the ultimate goal and work toward achieving it.

Sculpting

In ZBrush, the sculpting process begins. First, I create a quick basic blockout for the entire character. In this type of work, I always try to reuse base models from previous projects.

This approach saves time and prevents me from starting from scratch.

I apply polypaint as early as possible, even if the shape details are at a low resolution.
This helps me visualize the character as close as possible to the final look and achieve the desired impressions throughout all stages of work.

Afterward, I usually focus significantly on the character’s head. I avoid finalizing the sculpting concept until I’m completely satisfied with the resulting appearance.

Simultaneously, I create quick sketches of the head with various facial expressions. For the female fighter, I prepared two head variations with different smiles.

In ZBrush, I heavily utilize DynaMesh, especially during the blocking stages, and often continue using it until the final phase. Working with subdivisions on clean topology is reserved for the final stages.

etopology and Simple Rig for Posing

Since the character wasn’t intended for integration into a complex production pipeline, I didn’t have strict topology requirements. For this stage and similar tasks, automatic retopology in ZBrush followed by unwrapping in UV Master works well.

However, because I needed to create a simple rig in Maya later, I opted for a cleaner geometry approach for the body. I used a pre-made base mesh with good topology and transferred it onto my sculpture using Wrap 3D.

I applied the same mesh and approach for all characters in this series, saving a lot of time.

For the two heads, hair, and belt, I performed automatic retopology and UV mapping in ZBrush.

I kept the heads separate from the body mesh, allowing me to convey specific emotions for different angles and poses by toggling the visibility of the desired head in the rig.

Additionally, if needed, I can easily add a head with a new expression or even replace it entirely.

While this approach isn’t suitable for high-quality animation characters, it’s excellent for working with concepts, posing, and making quick changes without redoing the rig.

Creating rigs isn’t my strong suit, but basic rigging skills can be very helpful.

I used the Advanced Skeleton auto-rig for Maya, ngSkinTools for skinning, the Shapes plugin for minimal blend shapes (elbows, wrists, pelvic area, knees), and adjusted existing poses where the skin didn’t behave well.

Texturing

The texturing stage was very simple and efficient. I used polypaint from ZBrush to create diffuse textures and didn’t bake any additional maps. I focused on using color and tiny brush strokes to highlight accents and emphasize a painterly, drawn style.

For this character, I chose very light skin tones, characteristic of people with blonde or red hair. I also wanted each of the three characters to have their own color palette, and the skin tone needed to be different as well.

Additionally, the light skin tone is further highlighted by the black clothing.

Rendering

All characters in this series were rendered in Marmoset Toolbag 4. Setting up the scene was quite simple.

I imported the characters posed as needed and added a turntable operator.

The main focus here was on the lighting. I aimed to light the model in a way that avoided large shadow areas, preserved the purity of colors, especially in the shadows, and maintained the decorative effect of color and tonal spots.

At the same time, it was important not to lose the sense of volume. I used two-sided rim lighting, which I enhanced and exaggerated with warm and cool tones to further boost the decorative effect.

The material setup was very simple: connected diffuse textures, a high roughness level, and a small amount of subsurface scattering for smoother light and shadow transitions.

Next, I made minor adjustments to tone mapping, exposure, processing, camera position and angle, and export settings.

Post-Processing

After the renders from Marmoset were ready, I refined them using simple post-processing techniques.

The goal was to enhance the “drawn” and “painterly” effect of the images and add 2D effects like strokes and brush marks.

First, I imported the rendered frames into a 2D animation program.

In this case, it was TVPaint, which I find very convenient for drawing these kinds of effects, but any other frame-by-frame animation software could work. At this stage, I drew additional elements on each frame: strokes, brush marks, etc.

The next step was batch processing the frames with the Snap Art plugin.

It transforms images into various techniques (oil painting, pastel, watercolor, and many others). While it doesn’t always perform as well as I would like, it is sufficient to create an additional render pass.

Overlaying this pass on top of the main image adds more texture to the final picture.

Next, I batch processed the main set of frames in Lightroom, further adjusting colors, gamma, etc.

After that, I assembled everything in DaVinci, blending the layers and exporting the rotation video and individual frames.

I should note that this is not the only approach to rendering and processing that can be used. We could opt for other renderers with built-in stylized render pass options instead of Marmoset.

DaVinci, After Effects, or Nuke can also be used. To achieve 2D drawn effects, Toon Boom, Grease Pencil in Blender, or even AI tools will do the job.

Conclusion

And that is my method for working with 3D character concepts. It doesn’t include polishing, creating quality topology, textures, or a full rig for animations.

However, it seems very suitable for stages like pre-production. It allows for relatively quick visualization of ideas, checking the viability of the concept, and preparing blocking for further stages of production.

I hope this article will be helpful for those looking for similar methods and approaches. At the end here, I’ve included a bonus.

You can download the assets of this character via the link.

It’s free. Don’t forget to check out the description and terms of use. Thanks to the editors for inviting me to write this piece.

Wishing all readers inspiration and good luck in their creative endeavors!