

Loic Bramoulle
Hi, I’m Loïc Bramoullé, an art & film director from France. I’ve been working in game and animation studios in Paris since completing our graduation film at Melies School in 2010. Since 2015 I’ve been working mostly remotely as a freelancer, and in the last 3 years, I’ve been specializing in indie motion capture game cinematics with Blender.
Since I first tested iPhone facial mocap with CC3-based characters. My latest freelance cinematic gig was for the upcoming Mars Attracts game , which was nearly the most viewed trailer at Gamescom on IGN’s channel.
I used AccuRIG for the two first custom characters, CC4 and iClone to mocap the astronaut’s scream with my iPad.
Project HellGal
With HellGal I wanted to see how I could push my mocap cinematic pipeline to be as seamless as possible. Essentially merging mocap recording and film editing, to avoid all the time lost rendering individual shots and only seeing the final film at the last moment. The main thing I used for that is the under-exploited Scene Strips in Blender, allowing you to do non-destructive video editing on 3d scene(s), meaning you can select and time your animations to create shots, without actually cutting your animations.
All my mocap clips were imported on a big 5000 frames timeline, and just like in a video editor I edited my final cut from that, in 3d, without rendering these 5000 frames in any way. The big advantage is that when rendering, only my ~1000 final cut frames were considered, versus having to render a huge amount of frames and discarding them in post-edit.


Character Design – Inspiration and Concepting
Creatively, my goal was to design a female mashup of the DOOM Guy and Hellboy, set against the backdrop of a BLAME!-inspired ancient structure. This concept had been on my mind for years, and this contest provided the perfect excuse to dedicate as much time as I could. For HellGal’s design, I gathered references on Pinterest and began sketching, focusing on close-ups of her face to refine her character. For the environment, I opted for a rapid workflow, imagining it on the spot and kitbashing directly in Blender while iterating on the shots and animation as I went along.

Starting with Character Creator (CC)
I usually start with Kevin or Camilla to already have everything setup, realistic features, and wrinkles, but I just have to symmetrize the mesh in CC4. As they are scanned, it’s not ideal to sculpt an entirely new design with these slight symmetry differences. CC4 mesh editing makes it easy to symmetrize the face, then all design sculpt layers in ZBrush or Blender can be perfectly clean, and asymmetry can still be controlled independently with another sculpt layer.

ZBrush Face Tools for Character Creator is really the go to solution to design not only a character’s face, but its expressions and even wrinkles along with it. And that’s key for convincing facial animation, it’s not just about the motion data and design, but the actual custom expressions that really enhance the performance’s believability, instead of being generic and making the character feel more flat.
I worked on the design in a very iterative way, meaning I started with CC4 morphs to start customizing Camilla, then did most of it via FaceTools. Which allowed me to seamlessly preview her design and facial expressivity back in CC4. But I’ve also kept doing slight design variations in Blender while I rendered the various iterations of the full animation. So I could still polish her design while seeing her in the final shots, with the final lighting and cameras.



Wrinkle editing in Character Creator (CC)
The default Wrinkles packed in CC4 are already perfect in many cases, but I did want to stylize and paint them myself, to match the handcrafted style I was after. So after customizing a bit the sculpt of the wrinkles via FaceTools, and once HellGal was imported in Blender via the pipeline addon, I created a PSD file where I had the 4 albedo texture: the base face color, and the 3 wrinkles color textures.


Each time I painted over the face, I could update and refine the wrinkles simultaneously. By using Photoshop’s “File > Generate > Assets” feature, I enabled real-time saving for any layer ending with an extension (e.g., Layer_01.jpg). This meant I didn’t need to manually save textures—just open the PSD, paint, and let them auto-reload in Blender using the Auto-Reload addon. So I can essentially watch the final film, while seamlessly painting textures, her face and wrinkles for example.

3DCoat for Body Sculpt
I love 3DCoat for its ease of use compared to ZBrush, offering an intuitive workflow that feels almost as natural as working in virtual reality. Starting with voxels allows you to shape designs without constant remeshing, and the space mouse navigation is flawlessly implemented. Adjusting brush size and intensity is seamless, unlike ZBrush, where you need to open menus for similar tasks. While 3DCoat may not match ZBrush in terms of professional, production-level refinement, it excels as a creative tool, akin to a “3D Photoshop”. It also provides full asset production capabilities, including retopology, UV mapping, baking, and surfacing tools. Performance-wise, it holds its own, enabling real-time sculpting on models with tens of millions of polygons.



I essentially just used basic tools to sculpt the whole body and horns, move, clay, a few alphas, and had a “sketchy” approach, where I just made the armor like if I was drawing it on paper, or in clay. Professional character artists use extremely more complex methods in Zbursh, relying on an array of tools, blocking out forms, retopologizing, sculpting with carefully picked alphas. Here I just wanted to be expressive and design all the details directly by sculpting, in one single step pretty much. For the surfacing I decided to use Substance Painter instead of staying in 3DCoat, to get higher details and more pre-made materials, also as a challenge to use UDIMs to have very high fidelity while still dealing with only one single material and texture set.



Animation: VR with Mocap Fusion & iClone
After extensive research into the pros and cons of various motion capture solutions, VR mocap stood out as the most intriguing option. It’s an underexplored approach, not tied to a single product but a mix of potential setups—some outdated, others entirely untested. Despite its complexity, the appeal lies in its optical tracking capability, offering millimeter-accurate motion rather than the approximations of inertial suits. For instance, when properly configured, raw foot motion can be flawless, keeping feet grounded even on uneven surfaces. In contrast, inertial suits rely heavily on post-processing to generate such data since accelerometers lack positional awareness in 3D space.
Another significant advantage is its cost-effectiveness. For under $1,000, a setup with three Vive trackers and a Quest 2 can provide full-body tracking, including basic finger tracking. The system can then be expanded with additional trackers for elbows, knees, and other joints. Mocap Fusion (available on Steam) further enhances this setup by interpolating shoulder and elbow motion effectively, minimizing the typical IK snapping seen in character animations.

After months of research, I settled on a setup using Mocap Fusion with three Vive Tracker 3.0 units (for feet and hips), two Vive Base Station 2.0 units, and a Meta Quest Pro with controllers for hand tracking. To bring my virtual environment into the process, I imported a blockout of my 3D scene into Mocap Fusion, complete with collision data. This allowed me to act out scenes in my living room while interacting with the actual 3D geometry from my Blender environment. Best of all, I could embody HellGal’s character in real-time, seeing the final results at full scale as soon as the recording ended.
Animation editing in iClone
I exported the mocap takes as FBX files and imported them into iClone, where I had also loaded my environment blockout. This allowed me to select specific takes, refine their placement within the scene, and perform initial mocap cleanup using the Edit Motion Layer tool. The eye mocap, recorded along with the body in Mocap Fusion, was complemented by facial animation and dialogue captured in iClone using Live Face.
To finalize the workflow, I combined everything in Blender. This approach let me independently adjust the timing of body motion, facial animation, and audio while simultaneously refining the edit. By working this way, I could fine-tune shots, dialogue, and actions independently, ensuring they aligned seamlessly during the editing process.
Here are some iClone features I found particularly useful. Motion Direction Control helped accurately position mocap clips in space, while Motion Correction proved invaluable for fixing sliding feet and refining the character’s pose during foot contacts. This ensured a more polished and grounded animation.

One standout feature I utilized more extensively this time was AccuLips. It generates phoneme-based animations directly from dialogue audio and script, storing them in dedicated phoneme morphs or shapekeys. After fine-tuning these settings in iClone for a natural look, the entire AccuLips layer can be exported and edited further in Blender. In Blender’s graph editor, these phoneme animations remain separate from facial mocap shapekeys, allowing adjustments like smoothing or amplifying specific phonemes. This separation was especially helpful for achieving a more natural lip sync, particularly with tricky sounds like “B” and “P,” which iPhone-based mocap doesn’t always handle perfectly.
Blender Pipeline
I added all my animation exports from Mocap Fusion and iClone as strips in the NLA Editor, both for body and facial animations. This setup allowed me to edit each take individually and choose the best segments when actions were repeated during mocap. To refine shot composition and camera animation, I used camera markers, iterating directly within the scene. Between shots, I left ample empty space on the timeline and imported the 3D scene as a clip in the Video Sequence Editor (VSE) within a second scene in the same Blender file.
This approach enabled me to cut the timeline and retain only the portions of each shot I wanted, while simultaneously playing back the final cut. It also allowed iterative refinement across all aspects of the project—animation, cinematography, lighting, and editing—within a single file.


By integrating the entire pipeline into one workspace, I streamlined the process, avoiding the fragmentation and inefficiencies of traditional workflows. Instead of producing individual shots and combining them later based on a storyboard or animatic, this method felt more cohesive, like polishing a single, unified object. From character creation in CC4, mocap editing in iClone 8, to crafting cinematography and the final edit in Blender, the process became far more fluid and efficient. Here some really powerful addons I have tested for this project:
- HairTool: Easy to learn, with a robust library to kickstart any hairy or furry needs. Fully compatible with CC4.
- Wobbly Wiggler: Perfect for adding physics-based dynamics. Unlike most tools, it allows simulation not just for tails or ears but also for body bones like the spine, arms, and even fingers—enabling simultaneous animation and simulation.
- Octa Render: Soon out of beta, Octa Render has been a lifesaver for several projects. With rendering costs reduced by 60x compared to traditional methods, it makes high-quality animation both affordable and competitive with Unreal. Its simplicity is unmatched—a true one-click solution.
My Experience with Blender AutoSetup
The new AutoSetup significantly streamlines the process of sending characters and their animations into a Blender file with a single click. Compared to version 1, which required updates from GitHub, the new dedicated panel and button make the workflow far more intuitive and efficient. I plan to explore further how this setup supports seamless back-and-forth adjustments between CC4 and Blender, allowing for sculpting and customization of character proportions using Blender’s modeling and sculpting tools. This offers a strong alternative to relying solely on ZBrush with GoZ for detailed character modifications.


Here is a really good tutorial on using Blender as an alternative to Zbrush for customizing a character’s design and expression, and one-click synching back to CC4 with GoZ: https://www.youtube.com/watch?v=mgUTZ4GeItQ
The foundation of the character is built entirely in CC4, including the full facial expression set, the wrinkle system, body proportions, and even the Rigify auto-rig for final adjustments in the camera view. CC4 handles all the heavy technical work, allowing me to concentrate purely on artistic customization to align the character with my design vision.
My Thoughts and Suggestions
CC4 truly excels when users not only master its automated features to create characters that closely match their vision but also add a final layer of personal, handcrafted design. This could involve sculpting details in ZBrush or Blender and authoring textures in Photoshop or Substance, giving the character a unique and traditional touch. By developing a custom pipeline that integrates Reallusion tools into your workflow, you can fully own the creative process.


The seamless one-click integration of CC4/iC8 with industry-standard rendering apps like Unreal Engine, Unity, Blender, and Cinema 4D elevates it as a serious production tool. It eliminates countless tedious production steps, offering artists a significant technical advantage and streamlining the workflow.
Related Posts


























































































































