Hi! I’m Johannes Varmedal, Associate Art Director at Shapefarm, working out of our Tokyo office as Lead Environment and Tech Artist on Orbitals.
Orbitals is a two-player co-op space adventure game launching September 3, 2026 for Nintendo Switch 2. It features a retro-anime aesthetic that we have been deeply invested in crafting for the past three years in Unreal Engine 5.
Orbitals' unique world was envisioned by Marcos Ramos (Creative Director) and created by our incredibly talented art team lead Casper Wermuth (Art Director). Made as a love letter to 80s and 90s anime, Orbitals' art direction aims to capture the atmosphere of the era of anime that influenced us as artists—and to a certain extent, that originally brought us together as a studio.
Orbitals is a two-player co-op space adventure game launching September 3, 2026 for Nintendo Switch 2. It features a retro-anime aesthetic that we have been deeply invested in crafting for the past three years in Unreal Engine 5.
Orbitals' unique world was envisioned by Marcos Ramos (Creative Director) and created by our incredibly talented art team lead Casper Wermuth (Art Director). Made as a love letter to 80s and 90s anime, Orbitals' art direction aims to capture the atmosphere of the era of anime that influenced us as artists—and to a certain extent, that originally brought us together as a studio.
Unreal Engine’s source code access and flexibility enabled us to customize the default render pipeline and achieve our retro-anime art direction; one of the biggest technical challenges of the project. In pre-production, Thomas Söderlund (Technical Art) and Tony Holmsten (Concept Art) did the foundational R&D on how to achieve this.
The key observations in our research of the anime of the era pertained to the limitations and artifacts of its analog pipeline.
In development, this entailed creating custom shading models, adding our own render buffers, altering existing ones, and heavily relying on post-processing to achieve our target look.
The key observations in our research of the anime of the era pertained to the limitations and artifacts of its analog pipeline.
- Backgrounds at the time were typically painted with poster color on white paper per scene, with movement primarily limited to panning of the painting’s fixed perspective.
- Per frame moving components (characters / interactive objects / VFX) were commonly animated on twos at 24 frames per second, hand-painted on celluloid sheets (i.e. cels) typically behind pre-drawn linework, resulting in wobbling imprecisions in shading over frames.
- In compositing, cels were front-lit layered in front of the background, resulting in subtle separation or shadowing effects between the glass panes in the compositing stack.
- Emissive light effects were often achieved by back-lit holes punched into the background/cels for light to shine through, exceeding the brightness of unpainted paper.
- Each frame was ultimately a photo on film with the natural artifacts of the medium; grain, halation, chromatic aberration, bokeh, bloom, etc.
In development, this entailed creating custom shading models, adding our own render buffers, altering existing ones, and heavily relying on post-processing to achieve our target look.
Our cel shader
The shading models we created were split out in a fashion following the limitations listed above. The “front” layer in traditional anime compositing became our cel shading model–a technique frequently used in games–which steps shading into a limited amount of values, most commonly three. In our case, the three steps were set as per asset defined colors, utilizing our additional render buffers to give us full artistic control of our shading.
Taliesan Arnold (Lead Animator) spent a lot of time and effort refining the look and feel of our stepped frame character movement. VFX and all moving cel components that weren’t animated skeletal meshes were stepped in a similar way in our shaders, limiting movement to 12 fps (occasionally as low as 6 fps depending on effect) while our game camera rendering remained at a minimum 30 fps.
Our internal line rendering was defined with surface index fields and read into one of our customized buffers, then sampled together with the depth buffer in a common line creation UV-offset convolve setup to get both the internal and external lines as masks. The line mask result was further blurred to allow for alpha erosion differences in line thickness and noise-subtracted line breaks.
To emulate the imprecision in shading curvature and accuracy in lines over frames when anime artists painted their cel sheets, we noise-distorted our cel shading and line results.
To emulate the imprecision in shading curvature and accuracy in lines over frames when anime artists painted their cel sheets, we noise-distorted our cel shading and line results.
In-game, the wobble and line breaks are only updated when the cel actor or camera is moving; we thought of this as a “redraw” of the cel in traditional anime frame creation.
Focused on the analog hand-drawn nature of our references, although subtle, we found these kinds of imperfections important to the look and feel of our game.
We also added a highlight to our cel shadow step to add a common painting technique present in cels of the time.
Focused on the analog hand-drawn nature of our references, although subtle, we found these kinds of imperfections important to the look and feel of our game.
We also added a highlight to our cel shadow step to add a common painting technique present in cels of the time.
The final touch to our cel rendering was to add a subtle shadow artifact typically found in traditional front lit cel layer compositing.
No cel shaded actors ever cast computed shadows onto the environment as this would have contradicted the “front” cel and “back” environment difference in rendering.
To control this, we used a cel-unique lighting channel when lighting cel shaded actors, lit by cel unique light actors utilizing baked volumetric lightmaps.
To ground our characters and floor adjacent cel meshes, we simply used a cel-shaded shadow decal with an abstract shape.
To control this, we used a cel-unique lighting channel when lighting cel shaded actors, lit by cel unique light actors utilizing baked volumetric lightmaps.
To ground our characters and floor adjacent cel meshes, we simply used a cel-shaded shadow decal with an abstract shape.
Our environment shader
For the “background layer” (the fixed perspective painting in the case of anime), we defined this as our “paint shader”. Given our player-controlled camera, a fixed perspective limitation would be impossible to achieve; what we did go for was to define the rule that paint-shaded components and their lighting never move. The illusion we imagined the player accepting here was similar to our stepped movement for cel objects: that the player has effectively stepped inside the poster color painting, exploring the interior of the anime frame through the game camera.
Previously mentioned above for the lighting of cel-shaded objects, without any need for moving shadow casting, we opted to bake all lights as the expense of real-time lighting solutions would have been unjustified in our project.
We set up our paint shader to have full control of how surfaces are lit, the biggest contributing modifier being the ability to modulate light dispersion utilizing brush stroke textures.
We defined edge and corner highlights using vertex colors in our DCC source asset setup tool. In the editor, we controlled these using fake specular light source locators that could be controlled per actor in our editor tool.
The above, among other modifiers, entered the pipeline toolset that our art team used when creating assets to achieve a hand-painted “2D” feeling in our 3D environments. Our motto was to fight “digital perfection”: deliberately putting things off center; spacing details randomly to avoid repetition and symmetries; and distorting perfectly straight or round silhouettes.
The above, among other modifiers, entered the pipeline toolset that our art team used when creating assets to achieve a hand-painted “2D” feeling in our 3D environments. Our motto was to fight “digital perfection”: deliberately putting things off center; spacing details randomly to avoid repetition and symmetries; and distorting perfectly straight or round silhouettes.
Our film effects
Unreal Engine’s flexibility for customization meant that we could use our own post-process filters to allow for full control over the cel and environment rendering split in our pipeline.
We followed a similar approach in our use of bloom, which we masked so that the emitting surface remained cel or paint-shaded, while the bloom would surround it.
The final layer in our rendering was our film post-process filters. These included film grain, halation, chromatic aberration, and blur to simulate the feeling of our game frame being shot on film.
Our pipeline builds on Unreal Engine’s strengths, particularly its flexibility for customization. Beyond our shading model and render buffer changes, we relied on the Material Editor to develop and iterate most of our shaders directly in editor, keeping a tight link between artistic intent and the final result.
The pipeline was defined to enable the art team to be as expressive as possible within the constraints of our style. As a team already very familiar with Unreal Engine from previous projects, this enabled us to stay within established workflows while extending them where needed to support the visual direction of Orbitals.
To learn more about Orbitals, head over to the game’s official website.
The pipeline was defined to enable the art team to be as expressive as possible within the constraints of our style. As a team already very familiar with Unreal Engine from previous projects, this enabled us to stay within established workflows while extending them where needed to support the visual direction of Orbitals.
To learn more about Orbitals, head over to the game’s official website.