Chapter 1: Speed‑Cage Blockouts for Stance & Light

Created by Sarah Choi (prompt writer using ChatGPT)

Speed‑Cage Blockouts for Stance & Light

Unit 25 — 2D / 3D Hybrid Methods for Vehicles

Speed‑cage blockouts are fast, low‑fidelity 3D builds designed to answer two questions before you spend time on detail: how the vehicle stands on the ground and how it reads under light. The cage is a structural silhouette made from primitive volumes with honest proportions, set up to accept rough materials and simple lighting so the design’s stance, mass distribution, and value rhythm can be judged in minutes. For concept‑side artists, the cage buys rapid exploration and accurate paint‑overs. For production‑side artists, it anchors truth early—wheelbase, clearances, articulation envelopes—so later modeling doesn’t fight physics.

The value of a speed‑cage is that it delays polish while accelerating decisions. Stance is a composite of wheelbase, track, ride height, and rake; light is a composite of sun direction, fill strategy, and material response. A cage isolates these signals by replacing surface complexity with clean planes. When the cage looks good under a neutral HDRI and a single directional key, you know the eventual vehicle will survive optimization, different time‑of‑day lighting, and camera changes.

A practical speed‑cage starts with a brief, a scale reference, and a camera. Treat camera and ground as first‑class citizens. Lock the unit system and scene scale so one meter in software equals one meter in the world. Drop a ground plane with a subtle roughness and a calibrated gray so bounce light behaves predictably. Place a meter‑tall human proxy or seated driver rig to index proportions, then create a simple camera rig with a focal‑length range you expect from gameplay. If the title uses a fixed chase camera, match that lens and height to the centimeter; if the project is cinematic, build a couple of presets and label them explicitly.

Begin the cage with primitives: boxes for chassis and cargo, cylinders for wheels, capsules for struts, wedges for fenders, planes for aero. Snap pivots where real hinges and axles would live, not where modeling is convenient. Assign distinct material IDs even at this crude stage: one matte mid‑gray for the main body, a slightly glossier mid for glass proxy, a dark rubber for tires, and a bright metal for exposed mechanicals. These four reads are enough to test value separation without drowning in shaders. Keep bevels minimal or absent; sharp terminations make light behavior legible and reduce the temptation to polish.

Light the cage with a single directional key plus an HDRI or a broad fill, and only add a rim if your project leans cinematic. The key should be set at a believable sun elevation for your most common scenario; the fill should lift shadows without flattening the silhouette. Avoid exaggerated color early, and evaluate in both color and grayscale. If the silhouette collapses in grayscale, stance or proportion is under‑resolved, not the shader. Move the sun around the design, not the other way around, and look for angles where forms argue with each other. The speed‑cage is a form audit under moving light.

In a hybrid 2D/3D workflow, the cage is your paint‑over scaffold. Render a clean clay pass, an ambient occlusion pass, and a shadow‑only pass at high resolution with a linear color profile. In 2D, stack these as separate layers, lock their values, and sketch surfacing intent directly on top. Use multiply for occlusion, screen for subtle light lifts, and a neutral overlay for panel hints. Because the cage preserves faithful perspective and contact shadows, your paint‑over inherits credibility without the time cost of full modeling. When you find a better solution in paint, rebuild only the volumes that must change to keep 3D and 2D synchronized.

Kitbashing within a speed‑cage is about language and scale, not detail. Reuse neutral parts that carry believable affordances—generic hub units, standardized hinge blocks, radiator slabs, thruster cans, receiver rails—then decimate or boxify them so they behave like primitives in light. Downscale greeble density so the kit parts become value‑shaping planes rather than micro‑noise. The test for honest kitbash is whether the part continues the stance and reads at game camera distance; if it only looks good in a close crop, it does not belong in the cage. Keep a labeled library of “kit‑as‑primitive” assets with true measurements and neutral materials so you can drag and drop without re‑tuning.

Photobash ethics in the speed‑cage phase revolve around authorship, transformation, and truthfulness. Use photography to accelerate context, ground contact, sky reflections, and dirt logic; avoid lifting distinctive manufacturer signatures or unique concept art motifs. When images are used, document sources internally, transform them materially (perspective, lighting, color, silhouette integration), and avoid passing unmodified captures as design. A good rule is that a bash should never be the only evidence that a form exists; the cage must still stand on its own in a clay render. When working with external or community kits, respect licenses, note attribution where required, and avoid embedding assets that you cannot clear later for commercial production.

Stance testing should be ruthless. Spin the camera at crouch height, eye height, and top‑down metrics height. If the vehicle looks timid from low angles, widen track or lower the greenhouse; if it looks ungainly from above, redistribute mass by shifting cabin and power volumes. Use exaggerated ground planes to test reads: a 15‑degree ramp for approach and departure, a speed bump for breakover, a curb for sill height, or a landing pad grid for skids and gear. Animate a five‑frame suspension compression or gear deployment using just the primitive pivots; watch the silhouette under the same light as it squats, tilts, or plants. This micro‑rig reveals collisions and ugly poses long before there is a rigging team involved.

To keep the cage honest, maintain a small set of overlays. A silhouette mask helps you compare against earlier passes. A reflection‑only render checks gloss hierarchy and whether highlights fight the silhouette. A contact‑shadow pass verifies that the vehicle feels weighted rather than floating. Toggle these during reviews to keep feedback on stance and light rather than taste. When a change is requested, move a single governing dimension—wheelbase, track, cabin height—and re‑render before touching anything subtle. The cage is an instrument; tune only the strings that matter.

For production‑side artists, the speed‑cage is also the earliest packaging truth. Drop bounding boxes for engine, battery, fuel, radiator, turret baskets, or landing gear bays. Indicate keep‑out cones for sensors, recoil, or ingestion, and embed a simple human reach envelope. These volumes do not need edge fidelity; they need to be correct in size and position. Export a lightweight FBX or GLTF with named locators for these boxes so systems designers can begin simulation proxies while concept continues exploration. This two‑track workflow turns waiting time into learning time.

When moving from cage to polish, plan an intentional escalation. Replace primitives with minimal quads or NURBS patches only where the light needs curvature. Keep the rest planar until the design proves itself across a few lenses and times of day. As surfacing appears, lock a “cage freeze” snapshot so you can always return to stance if curves start to wobble the read. Resist tertiary detail until materials and value composition are holding up. At any hint of drift, render the old cage and new surface under the same lighting and overlay them in 2D. If the delta is hurting stance, roll back.

File hygiene keeps the method fast. Name everything by function rather than look: chassis, cabin, power, intake, exhaust, gear.FL, gear.FR, wheel.RR, ramp, hatch, gimbal, sensor. Keep transforms frozen and pivots meaningful; speed‑cage rigs decay quickly if transforms are dirty. Save camera presets with lens, height, and tilt annotated in the name. Maintain a simple lighting rig file that you never paint‑over directly; always render out to new plates to preserve your benchmark.

A compact case study illustrates the approach. A studio needs a compact courier skimmer for dense urban lanes. The speed‑cage begins with a 3.0‑meter wheelbase equivalent defined by thruster spacing, a low 1.2‑meter body height, and a forward 10‑degree rake for urgency. Cylinders stand in for ducted fans, a box defines cargo, and a capsule marks the pilot. A neutral HDRI and a 35‑degree directional key define the light. From low angles, the skimmer feels tall and narrow, so track widens and the cargo drops by 80 millimeters; the silhouette now plants. A quick five‑frame “bank and brake” pivot reveals the tail cone clips a sensor dome; the dome slides forward and the tail planes thicken, which also improves highlight flow. The painter takes a clay, AO, and shadow pass into 2D, adds a high‑contrast livery and a luminous strip, and draws a sharper shoulder line. The 3D artist replaces the body box with two lofted patches and keeps everything else planar. The team freezes the cage as a benchmark, proceeds to kitbash generic hinge blocks, and documents one photo texture used purely for floor scuffs with source and transformation notes. The vehicle ships with the same stance and light read the cage promised.

The ethics of speed are also the ethics of collaboration. A speed‑cage makes strong claims early, so it must keep receipts: scale bars on renders, camera notes in filenames, unit declarations on plates, and short paragraphs that explain intent and risks. When concept and production both treat the cage as a shared instrument, the studio gets the best of both worlds: bold exploration that never strays from buildable truth. The result is a vehicle that plants on the ground, sings under light, and moves through the pipeline with momentum rather than friction.