Chapter 4: Safety Devices & Warnings as Diegetic UI

Created by Sarah Choi (prompt writer using ChatGPT)

Safety Devices & Warnings as Diegetic UI (No‑Bypass Guidance)

Safety devices and warnings are not decorations; they are the prop’s native user interface. When built into silhouette, proportion, and surface logic, they guide behavior without pop‑ups or subtitles and reduce the temptation to bypass. For prop concept artists, the task is to encode safe operation, hazard awareness, and state transitions as visual, tactile, and audible cues that survive the game camera. This article frames safety as diegetic UI, showing how to design interlocks, guards, and warnings that read instantly, remain coherent across states, and carry through to production.

A strong safety design answers five questions in half a second: where is it safe to touch, what is about to move, which parts are hot, pressurized, sharp, or energized, what must happen before the action, and what will confirm that the action occurred. The silhouette must broadcast these answers before materials complete the story. In production, those answers must persist when poly counts drop and when the prop shrinks on screen.

A safety affordance begins with proportion honesty. If the user must grip near a hazard, the handle should dominate with a protective lip or guard that is visibly thicker than surrounding skins. If a cover shields a blade or pinch point, its hinge and sweep path need to be evident so the viewer can predict the motion. A believable safety interlock looks like it would mechanically block misuse; a believable warning looks like it would be visible even in poor light. These are shape problems first, texture problems second.

The grammar of diegetic warnings relies on families of form. Rounded volumes and broad fillets communicate safe contact, while triangular wedges, chevron bites, and proud fins communicate direction and danger. A safety catch that must be lifted before a lever can slide should sit perpendicular to the main travel, with just enough offset that the sequence is self‑evident: lift, then pull. The state change should alter the silhouette in a small but readable way, such as exposing a red tag when the guard is open or revealing a slot when a pin is withdrawn. The more critical the action, the larger the silhouette delta between states.

Interlocks discourage bypass when their utility is visible. A bayonet‑style latch that keys into a notch, a hasp whose body clearly blocks the lever path, or a shutter that intrudes into the danger zone makes the viewer intuit that defeating it would require tools or risk. If the fiction allows electronics, the physical constraint should still tell the story alone; lights and audio then become reinforcements rather than crutches. When depicting, choose perspectives that reveal the interfering geometry and avoid angles that hide the very feature that enforces safety.

Warnings work best as layered modalities. The first read is shape and placement: proud guards and halos around hazard zones, offset planes that cast protective shadows, and negative spaces separating safe and unsafe touch areas. The second read is value and material: matte elastomer for grips, satin paints for covers, dull darks for interior cavities that should not be touched. The third read is pattern and iconography, which should be restrained and scale‑aware. If a warning stripe will collapse to two pixels at gameplay distance, the silhouette must already be doing the work. Texture should never carry meaning alone.

States must be obvious at a glance. A safe state should look physically calm: guards closed, handles retracted, and axes parallel to the housing. A ready state might reveal armed geometry: a spring compressed, a pin raised, a shutter rotated into alignment with an indicator. A fault state should trigger a shape inversion: a flag pops, a tab protrudes, or a window reveals a contrasting block. These transformations should be staged so that even a blurred camera pan registers the change. If a prop has multiple independent hazards—heat, pressure, radiation—each should have a distinct silhouette cue so one does not masquerade as another.

Motion cues prevent hands from wandering into the wrong place. A lever’s long arm should extend away from the pinch region so that pulling naturally moves fingers clear. A lid that opens into a hot cavity should arc away from the user’s face and present its hot surfaces inward. Rotary tools should show spin trains with stepped guards that anticipate fling paths, while vents and louvers should lean with airflow to predict blast direction. Place these cues on planes that face the likely camera so they are not lost to occlusion.

No‑bypass guidance depends on visible sequencing. The viewer should be able to read the order of operations with no text: remove the safety pin, rotate the guard, then pull the primary lever. Designing for this means giving each step a different axis, a different silhouette notch, and a different tactile signature. The pin should be orthogonal to the lever; the guard should swing rather than slide if the lever slides. Each motion should expose or conceal the next affordance so that skipping a step feels physically blocked. In concept sheets, show the steps as three silhouettes with simple arrowed callouts to lock the mental model.

Accessibility is part of safety and should be designed, not bolted on. Shape cues should remain legible to players with color‑vision deficiency; value contrast and geometry must carry the message. Tactile cues like scallops, ribs, and flat‑to‑round transitions should be visible enough that their presence implies feel. Audio cues should not be the only confirmation; a detent window, a spring‑relief gap, or a flag should provide visual closure. If left‑handed operation is plausible, symmetric guards and mirrored affordances prevent accidental crossing into hazards.

Cinematic clarity should not undermine credibility. Low camera angles can make a guard look thinner than intended, so thicken silhouette‑critical lips and add a second step at the edge so it holds a highlight at distance. If a hinge is tiny in orthos but must be read in a hero shot, enlarge the knuckle and exaggerate the spacing between leaves; the model can swallow some realism if it protects communication. Conversely, avoid gratuitous “safety jewelry” that reads as cosplay. Each safety detail must have a plausible role and mechanical logic.

Production survivability begins with annotated metrics. Minimum edge thicknesses for blades and fins, minimum wall widths around finger voids, and minimum radii for grip fillets should be called out so decimation does not erase safety intent. LOD plans should preserve guards, stops, and pins as separate silhouette elements to LOD2; if they merge too soon, the diegetic UI disappears and the player loses trust. Bake highlights on proud safety planes and avoid relying on micro‑normal stripes that will vanish in MIP maps. When possible, test a blockout in engine to verify that rim light or emissive edges are not needed to understand safe vs. unsafe.

Failure patterns tend to repeat. Props that signal danger only with color and decals become toothless when desaturated by distance or lighting; their silhouettes must be re‑sculpted to include bites, guards, or proud stops. Props that hide interlocks on the back side fail when animated; their sequencing needs to be re‑oriented so the camera sees the block. Props that rely on flush buttons for critical actions invite mis‑press; the actuator should be recessed or covered with a flip guard if accidental activation would be catastrophic. These fixes are form‑first and should be solved in the black‑shape stage before materials distract.

Collaboration between concept and production keeps safety honest. Concept artists should document non‑negotiables such as “guard must occlude 30% of blade in stowed state,” “safety pin must protrude 6–10 px at gameplay distance,” or “lever cannot travel unless tab is lifted 5 mm.” Production should mirror these in model notes and validate with low‑poly rigs. When rigging or collision volumes demand changes, the team should choose the solution that preserves the no‑bypass story even if it costs a few more triangles where it matters.

Testing is simple and ruthless. Shrink the sheet to target on‑screen size and ask a teammate to describe how to operate safely. If they are unsure, the silhouette has not earned trust. Blur and play a slow pan; if the state change fails to pop, increase the silhouette delta. Flip the image horizontally to catch asymmetric hazards that bias against left‑handed reads. Place the prop against both bright and dark backgrounds; the safety cues must survive either.

Ethics anchors the whole effort. Safety depicted without consistency teaches players the wrong lessons. If narrative demands a reckless device, signal that recklessness honestly with sloppy guards and improvised warnings so the world still owns the consequences. If a prop is ceremonial but dangerous, resist hiding hazard cues behind ornament; let the language of risk remain legible so that accidents feel diegetically true rather than designer‑induced unfairness.

A practical workflow starts with silhouettes that already include guards, interlocks, and warning deltas between states. Progress to gesture studies that exaggerate the arcs of safe motion. Lock anthropometric clearances and minimum thicknesses as numeric callouts. Choose a hero perspective that shows interference geometry and the order of operations. Test in engine as early as possible to confirm read at distance. Treat the final material pass as a way to reinforce what shape already says, not as a crutch for what it does not.

Safety devices and warnings, when treated as diegetic UI, make props self‑explanatory and trustworthy. By embedding no‑bypass guidance into form, you ensure that players read the right action, at the right time, in the right place—without ever needing a tooltip to keep their hands safe.