Sarah Clarity Studios

Chapter 1: Cage Blockouts & Paintovers

Created by Sarah Choi (prompt writer using ChatGPT)

Cage Blockouts & Paintovers for Mecha

Cage blockouts and paintovers sit at the center of 2D/3D hybrid workflows for mecha because they solve a hard problem: getting believable structure, perspective, and functional clarity quickly—without locking you into a final design too early. A “cage” is a simplified 3D scaffold: a volume map of the mech that establishes scale, proportions, joint placement, and big-plane orientation. The paintover is where you turn that scaffold into design—shape language, panel logic, materials, wear, faction cues, and story.

This article is written for both sides of the mecha pipeline. If you’re on the concepting side, cages help you iterate faster and make your silhouettes and orthos more consistent. If you’re on the production side, cages reduce rework by making form, clearances, and intent explicit, and they become a shared reference for modeling, rigging, VFX, and animation.

Why “cage” thinking matters in mecha

Mecha designs fail more often from structural ambiguity than from lack of style. A cage forces decisions early: where mass lives, how limbs clear armor, where the cockpit and power core are, how feet contact ground, how weapons are braced, and how the center of mass behaves in a stance. Even in stylized work, the cage gives you a consistent internal physics. That consistency is what lets you exaggerate confidently.

A cage also creates a common language between artists. A strong 2D design can still collapse when it’s turned, posed, or animated. The cage is a cheap “truth check.” When you rotate the blockout or drop it into an action pose, you immediately see where the design cheats break. Those failures are valuable; they tell you which areas need redesign before you invest in finals.

When to use cage blockouts vs straight 2D

You don’t need 3D for everything. Use cage blockouts when you need one or more of these: multiple views that must match (orthos, callouts), a complex pose or camera angle, believable transformation sequencing, tight hard-surface layering, interlocking armor, complex weapon mounts, or clearances for animation and rigging. Cage blockouts are also worth it when the mech will be reused across skins, variants, or a family lineup, because the shared scaffold becomes a proportional anchor.

Stay mostly in 2D when you’re still exploring silhouette families and you need maximum ideation speed. Many artists overcommit to 3D too early and “freeze” the design. A good rule is to keep your first 30–60 explorations in pure 2D (silhouette banks, quick line ideation), then move into a cage once you have a promising direction and you need to test it from multiple angles.

The hybrid pipeline at a glance

A practical 2D/3D hybrid loop for mecha often looks like this: brief and constraints → silhouette ideation → cage blockout → paintover iteration → refinement sculpt or hard-surface pass → callouts and orthos → production handoff package. The key is that you cycle between 3D and 2D in small loops instead of doing “all 3D first” or “all 2D first.” The cage is there to keep your loops honest.

On the concepting side, your loops answer design questions: “Does the shoulder read as heavy armor or a weapon bay?” “Does the leg feel like it can carry that torso?” “Can the head/visor be seen in gameplay camera?” On the production side, the loops answer implementation questions: “Can the rig hit this pose without intersections?” “Is there room for recoil travel?” “Where do VFX and audio events live?”

Building the cage: volumes first, not details

A cage blockout should begin as big primitives: boxes, cylinders, wedges, and simple curves that establish proportion and gesture. Think of it as a mannequin for a mech. Keep the silhouette readable and the planes clean. If you start with greebles, you will bury the design decisions under noise and you will also slow yourself down.

Start by placing the major mass groups: head/sensor block, chest/core, pelvis/hip block, upper legs, lower legs, feet, upper arms, forearms, hands or tools, and any major backpack, thruster, or wing masses. Then lock the “hinge logic” with placeholder joints: shoulder pivots, elbow pivots, wrist pivots, hip pivots, knee pivots, ankle pivots. Even if the mech is non-humanoid, you still need pivot logic, because pivots define animation reality.

Once volumes are placed, establish the mecha’s “skeleton line.” This is a simple gesture curve that runs through the stance. A cage without gesture becomes a stiff statue, and your paintovers will fight that stiffness. Good cages have posture—weight shifted, stance grounded, and readable balance.

Scale, camera, and the gameplay read

Mecha design is always about scale, even when you’re not drawing a size chart. Your cage blockout is the fastest way to keep scale believable. Drop a human proxy next to the mech, then test the mech at the expected gameplay camera distance. Many cool cockpit ideas become unreadable or implausible once you place a pilot, canopy thickness, and access ladder.

For concepting, this helps you design “hero reads” that survive in gameplay: head shape, shoulder silhouette, main weapon silhouette, and major color blocking. For production, this prevents the classic failure of micro-detail that never shows and eats texture budget. A cage encourages you to prioritize big shapes and only add detail where it supports recognition.

Clear pivots and collision: the production-friendly cage

A cage becomes production-friendly when it includes clearance thinking. Shoulder armor must clear the torso in a raise. Knee plates must clear in full crouch. Backpack thrusters must not intersect the head in a look-up. Weapon barrels must have recoil travel space. Hands need grip clearance on tools.

In a cage, you can test these quickly by rotating parts and pushing to extremes. You don’t need a full rig; you need “range-of-motion sanity.” Mark collision zones with simple colored volumes or overlays in your screenshots. Production artists will thank you because you are showing not just what the mech looks like, but what it can do.

From cage to paintover: controlling what 2D is for

A paintover should not be a “cosmetic skin” on top of random 3D. The 3D should handle perspective and structure; the 2D should handle design decisions and taste. When that division is clear, you work faster and your results look intentional.

Start paintovers by clarifying primary, secondary, and tertiary shapes. Primary shapes define silhouette and role. Secondary shapes define armor segmentation and function groupings. Tertiary shapes are surface detail and wear. If you skip to tertiary, your design will look busy but unclear.

In paintovers, keep a consistent light direction for readability, then use value grouping to separate masses. Mecha readability improves dramatically when you simplify values into “big families” rather than shading every plate individually. After values are grouped, apply material cues (painted metal vs bare metal vs rubber vs glass), then add decals, warnings, and story wear.

Kitbash: using libraries without losing authorship

Kitbashing is one of the fastest ways to build cages, especially when your studio already has a kit library. It’s also risky: it can trap you into the kit’s design language and make your work feel generic. The best kitbash use is structural, not stylistic. Use kit parts to establish plausible forms—actuators, joint housings, pistons, vents, thruster bells—then redesign the silhouette and primary shapes so the mech still feels authored.

A strong method is to kitbash only the “engineering layer” first: joints, pistons, hoses, and major mechanical supports. Then design the “armor layer” as your authored shapes. This gives you believable under-structure without surrendering the design.

For concepting-side artists, kitbash can accelerate exploration across variants, but you should still do 30% deltas: change the big silhouette and role read between variants, not just the surface detail. For production-side artists, kitbash can align your concept with what the studio can actually build, because you’re borrowing proven proportions and part language.

Sculpts and hard-surface refinement: when the cage graduates

Sometimes the cage stays simple through the whole concept phase. Sometimes it needs to “graduate” into a more refined 3D pass: a quick sculpt pass for silhouette and plane control, or a hard-surface pass for panel breaks and manufacturing logic.

A sculpt pass is helpful when your mech relies on strong plane design and you want clean lighting reads in renders. It’s also good for stylized mecha with bold, simplified forms. A hard-surface pass is helpful when the mech has layered armor, complex vents, and precise mechanical interfaces that must be consistent across views.

The main danger is spending too long in 3D refinement before design decisions are stable. Treat 3D refinement as a tool to answer a specific question: “Do these armor layers actually stack?” “Does this knee mechanism clear?” “Does the backpack silhouette read in profile?” If you can’t name the question, you might be polishing too early.

The screenshot kit: outputs that speed up paintovers

A good cage workflow produces a small set of screenshots that become a paintover kit. Typical essentials include a 3/4 view, front, side, back, and a dynamic pose. For complex mecha, add a top view for backpack/shoulder layouts, and a low-angle view for intimidation reads.

In addition, generate “line renders” or clay renders that emphasize form without material noise. These are ideal underlays. If you can, render an AO pass and a simple shadow pass. You can combine these in your paintover stack to get believable grounding quickly, then paint design decisions on top.

For production, these renders can be included in handoff packages as “intent frames,” showing how the mech should look under consistent lighting.

Photobash and photo-texture ethics for mecha

Photobash is a powerful accelerator, but it comes with ethical and legal responsibility. The safest approach is to use your own photos, studio-provided libraries, or properly licensed stock and textures. Avoid using copyrighted images that you don’t have rights to, especially recognizable product photos, movie stills, or other artists’ concept art, because those can create legal risk and can also undermine trust in your authorship.

In practice, you can use photos in two ways. The first is “reference extraction,” where you study a photo and repaint it so no original pixels remain. The second is “texture collage,” where you integrate photo textures into your concept. Texture collage is where licensing matters most, and it’s also where your work can accidentally inherit real-world branding, serial numbers, or protected designs.

If you’re on the concepting side, ask what your studio’s policy is for photobash and stock use, and follow it strictly. If you’re on the production side, you may be receiving concept art that uses photo elements; treat those as concept-only and don’t assume they can ship. A clean handoff includes a note that specifies whether any photo textures were used, and whether they are licensed for production.

Clean authorship: making hybrid work feel like “you”

A common fear is that 3D makes everything look the same. The cure is to establish a design system before you open your 3D scene. Decide your silhouette family, your proportion ratios, your panel language (long bands vs mosaic plates), your edge philosophy (hard chamfers vs soft bevels), and your detail density rules.

Then, in paintovers, emphasize your authored primary shapes first. Use 3D for truth, but use 2D for taste. When you do this, the 3D becomes invisible, and what the viewer sees is a strong design.

Common failure modes and how to avoid them

One failure mode is over-detailing the cage. If your blockout already contains small greebles, your paintover becomes confusing and slow. Keep the cage at the “big planes and pivots” level until you have a locked direction.

Another failure mode is using kitbash as a crutch. If the mech’s silhouette looks like the kit set, your authorship is weak. Counter this by redesigning the armor layer and doing silhouette checks at every iteration.

A third failure mode is perspective inconsistency across views. The cage solves this only if you actually use it as the source of truth. If you paint over one view and then freehand the orthos, your final package will drift. Instead, produce all views from the same cage, then paint over them with shared layer groups.

Finally, a big failure is ignoring production constraints. If your design has impossible overlaps, zero clearance, or no space for rigging, you’ll create late-stage pain. Use the cage to test range-of-motion early and annotate collision zones.

Handoff: what to deliver from a cage + paintover workflow

For concepting-side deliverables, include the cage-based views and the paintovers, plus a small page of key callouts: joints and pivots, weapon mounts, sensor suite, cockpit access, and any transform or deploy sequences. Include at least one “clean read” render that shows the design without heavy effects.

For production-side deliverables, add more explicit structure: orthos derived from the cage, pivot markers, notes on clearance assumptions, and a materials map or guide. If you used kitbash, list the kit sources if they are internal assets, and clarify whether any third-party assets were used and what their licensing status is. If photobash textures exist, flag them as concept-only unless they are confirmed shippable.

A final best practice is to include the cage file or an exported proxy mesh (if studio policy allows). Even a simplified proxy can save modeling time and preserve the intent of your proportions.

A mindset that keeps hybrid workflows healthy

Cage blockouts and paintovers are not about “cheating” or replacing drawing. They are about choosing the right tool for the right problem. Use 3D to protect structure and save time. Use 2D to design, communicate, and bring taste. When both sides work together, you get mecha designs that look cool, turn cleanly, animate believably, and hand off smoothly to production.

When in doubt, return to the core question the cage exists to answer: “Does this mech work from every angle, in motion, at gameplay camera distance?” If your cage and paintover workflow keeps answering that question, you’re using the hybrid method the way it’s meant to be used.