Chapter 3: Edge Control & Brush Economies

Created by Sarah Choi (prompt writer using ChatGPT)

Edge Control & Brush Economies — Style Systems: Stylized ↔ Realistic

Why Edge Control is a Style Engine

Edge control is where form meets perception. Before textures or shaders, the way you transition between values and shapes tells the viewer what the object is made of, how it moves, and how important it is. In stylized projects, edge decisions are bold and systematized—long clean silhouettes, intentional lost‑and‑found edges, and minimal texture noise. In realistic projects, edge behavior is more granular—micro bevels, occlusion wedges, and material‑specific highlight sizes—but it still follows rules so the cast feels cohesive. Brush economy is the practical counterpart: a constrained, well‑labeled brush kit that produces consistent edge behavior across concept paint, callouts, and marketing art, making PBR translation predictable for production.

The Stylization Continuum for Edges

Think of edge design as a slider like proportion and material. On the stylized end, you compress edge varieties to a few families—clean graphic cuts, soft organic tapers, and selective texture breaks. On the realistic end, you expand the catalog—crisp plane breaks at bony landmarks, rolled edges on leather, feathered cloth transitions, sharp metal chamfers, and glancing Fresnel falloff. Most shipped styles mix a constrained silhouette language with selective realistic micro‑edges in focal areas. Document where each family is allowed and at what frequency, then maintain it from blue‑sky sketches to final orthos.

Shape and Edge: One Story

Edges exist to explain shape. Large, simple masses deserve long, stable edges with sparing interruptions; small accessory forms can carry higher edge frequency. If your tank silhouette is built from megashapes, resist peppering the silhouette with tiny notches that dilute the read. Conversely, a striker reads faster with longer tapered edges and pointed terminations that imply speed. Lock three decisions early: which forms own straight segments, which own S‑curves, and where transitions must be staged (hard→soft through a middle‑value band) to keep volumes turning without mush.

Value Structure and Edge Priority

Value and edge are inseparable. A “hard” edge is often just a high contrast meeting with minimal transition pixels; a “soft” edge is a graded handoff that compresses contrast. In stylized modes, you group values into two or three families and let edge placement do the speaking; in realistic modes, you let subtle value shifts describe plane change while preserving accent edges at contact and specular zones. Establish a contrast hierarchy by role: head/hands/insignia earn the sharpest value breaks; torsos carry medium breaks; legs and capes accept softer transitions so motion blur doesn’t strobe. If a grayscale thumbnail at 128 px reads with three steps—silhouette, mid‑planes, accents—your edge plan is working.

Palette Discipline and Edge Perception

Color temperature modulates perceived edge sharpness. Warm next to cool reads sharper than warm next to warm at equal value. Use this to concentrate “virtual sharpness” near focus without increasing value contrast everywhere. In stylized worlds, reserve complementary jumps (e.g., teal vs orange) for facial features, emblems, or weapon glyphs; in realistic worlds, confine large temperature shifts to material boundaries (skin→metal, cloth→chrome). Document HSV bands per family so colorists and texture artists don’t accidentally soften a focal edge with equal‑temperature neighbors.

Brush Economies: Limited Tools, Infinite Outcomes

A brush economy is a curated set of brushes with explicit roles, named and versioned so teams can reproduce looks. Fewer brushes mean stronger style memory across the project. A robust kit typically includes: a hard‑round with controlled falloff for graphic cuts; a soft‑round for value staging; a square/tilt brush for planar hatching; a texture stamp for cloth/leather fibers; a rake for hair and brushed metals; and a smudge/blur tuned for edge welding rather than smearing. Each brush should have a documented edge profile at multiple sizes and opacities. Publish a short PDF or canvas swatch: “Brush 01—Hard Cut: no scattering, 0% transfer jitter; Brush 02—Chisel: angle‑tilt responsive; Brush 03—Velvet: soft with low‑frequency noise for cloth; Brush 04—Weld: smudge 8–12% for fusing values along form.”

Stroke Taxonomy and Edge Recipes

Pair brush metrics with stroke recipes. For example, “Graphic Cut” = single confident stroke, 95–100% opacity, no jitter, used only on silhouette or hero insignia; “Turned Form” = two‑step ramp: midtone block‑in → soft weld with 10–20% opacity; “Specular Kiss” = narrow, high‑value stroke with width tapered by pressure, only within highlight ellipse; “Lost Edge” = lift hue into background temperature and reduce contrast by 20–40%. Name these recipes and call them out on paintovers so others can replicate them. The taxonomy should also track frequency limits: no more than one specular kiss per 30° of silhouette arc for stylized metals, for instance.

Edge Families by Material

Edges inherit from materials. Skin needs rolled transitions around fatty pads and tighter breaks at cartilage (ala alar rim, tragus). Leather holds a mid‑hard break on cut edges with subtle darkening from burnish; cloth carries feathered edges and shifted saturation along fold crests; painted metal carries crisp chamfers with occasional micro‑nicks only at contact points; polished metal shows razor‑hard specular edges but soft albedo transitions; hair uses clustered “bundle edges” where clumps meet and single‑strand flyaways only at focal zones. Write these as re‑usable paragraphs on the sheet so texturing can author roughness and normal maps to match.

Silhouette Discipline and Edge Frequency

Silhouette is the loudest edge. In stylized systems, keep the silhouette frequency low and the interior edges doing the storytelling. In realistic systems, silhouettes can carry more secondary forms, but still obey scale: large body masses get long arcs; prop attachments inject mid‑scale corners only where functionally justified (hinges, vents). Always test silhouettes on neutral backgrounds and in the target camera height. If your design dies at 128 px, reduce silhouette frequency by 20–40% and move interest inside with value accents.

Lost‑and‑Found Strategy

Losing edges is not a mistake; it’s a compositional tool. Use lost edges to push depth hierarchy and to merge low‑priority forms into the value group behind them. Regain the edge at strategic beats—near the face, hands, weapon grip, UI‑tracked components. In stylized art, you can even “teleport” edges across a cape or skirt by carrying only the cast shadow and letting the form edge dissolve. In realistic art, let ambient occlusion and reflected light do the work: occlusion tightens the base of a form; bounced light recovers the far edge without a line.

Camera, Distance, and Motion Blur

Edge sharpness collapses with distance and motion; plan for it. For FPP, you can maintain higher edge contrast and sharper specular around hands and weapons; for TPP, lift mid‑tone compression and soften transitions on legs and capes to prevent shimmer. Under TAA, tiny alternating hard/soft patterns create crawling; design edges with a minimum on‑screen pixel width and avoid zebra striping in normals. Provide a “distance audit” strip on the sheet: full body at gameplay size, medium crop, and close‑up, each showing how edges are simplified.

Production Translation: From Brush to PBR

Every brush choice becomes a shader or a bake. Hard‑round graphic cuts imply crisp geometry or well‑defined chamfers in the high poly; soft welding suggests larger bevel radii or roughness gradients; texture stamps cue micro‑normal frequency and orientation. Annotate the intent: “graphic cut = 2–3 mm chamfer; velvet blend = roughness 0.65→0.55 over 25 mm; brushed metal rake = anisotropy direction map along local X.” This avoids the common failure where beautiful paintovers become noisy, over‑textured models.

Team Interop: Concepting vs Production

Concept‑side artists should establish edge policies in the first week of a style test: silhouette frequency caps, allowed lost‑edge zones, and three material edge recipes per family. Provide clean paintovers with edge notes instead of relying on ambiguous texture noise. Production‑side artists should enforce numeric rails in bakes and textures: bevel denominators, roughness bands, AO intensity caps. Tech art should test edges under the game LUT and motion to catch shimmer early, and shader authors should expose controls for highlight width so marketing doesn’t stray from in‑engine truth.

Common Pitfalls and Fixes

The most frequent issue is “edge salad”: mixed edge families applied randomly. Fix by limiting the brush kit and writing an edge plan per character. Another is “contrast everywhere”: when every edge is sharp, nothing is; reduce 20–30% of mid‑tier edges to soft or lost. “Texture fighting form” happens when high‑frequency stamps cross silhouettes—reserve texture to interior planes and align normal detail with flow. Finally, “specular glitter” occurs when micro‑normal noise is too high; lower amplitude or increase roughness and keep highlight shapes stable.

Practice Drills

Do a two‑hour “edge sprint” with a single hard‑round and single soft‑round brush—no texture. Paint the same bust three ways: graphic stylized, median, and realistic, changing only edge handling. Next, create a brush swatch sheet showing your kit’s edge profiles at multiple sizes and opacities. Finish with a distance audit at 128 px, 512 px, and 1,200 px; mark which edges survive and which collapse, then adjust your plan accordingly.

Final Thought

Edge control and brush economy are production tools disguised as aesthetics. When shape, edge, value, and palette rules align—and your brush kit encodes those rules—you get characters that read instantly, animate cleanly, and translate into PBR with minimal friction across the stylized ↔ realistic spectrum.