Designing 2D location architecture for games

2D Location Architecture Design

"Architecture of a 2D location" is not placing decorative elements. It's designing space where gameplay will happen: where colliders sit, how depth layers (parallax) are structured, how tiles connect without artifacts, how the player reads movement direction and danger zones without additional hints.

Visual and function are inseparable in a 2D location. A decoration looking like a platform — the player will try jumping on it. Danger zone without color coding — the player will step on it and be surprised.

Layering: the most technical part

Modern 2D location consists of many rendering layers. In Unity this is Sorting Layers + Order in Layer, in Godot — CanvasLayer with z-indices. Typical structure:

• Background (BG): static sky, distant mountains — minimal parallax or static
• Far Parallax: slowly moving far objects (0.2–0.4× camera speed)
• Mid Parallax: middle ground (0.6–0.7×)
• World Layer: game space — platforms, floor, walls with colliders
• Character Layer: characters, NPCs, enemies
• Foreground: decorations in front of character (0 parallax or 1.1–1.3× for reverse effect)
• UI: interface on top

Layer order mistake — character renders behind foreground trees though should be in front. Fixes in 30 seconds in settings, but often discovered only in final level test.

Tilemap architecture and AutoTile

For tilemap locations, key question is tile-connection rules. Unity TilemapRule (Rule Tile) automatically selects needed tile based on neighboring tile presence of same type. Properly configured Rule Tile eliminates visual artifacts at joints — inner corners, transitions between surface types.

Common problem: tiles drawn with wrong sizes. Basic tile 16×16 px at scale 1.0 should give exactly 1 unit in Unity (if 16 PPU — pixels per unit). Mixing tiles with different PPU in one scene gives "floating" objects — they don't connect though visually should.

Collision tiles vs visual tiles. In complex locations colliders shouldn't exactly match visual tiles. Grass on top of platform — visual element, collider runs along bottom edge of grass layer, not top. This gives feeling that character stands "on grass," not "on box." TilemapCollider2D + CompositeCollider2D combine tile colliders into one mesh — important for performance with many tiles.

Visual navigation and readability

Player should understand in 1–2 seconds in location: where you can walk, where dangerous, where to go next. This is location architecture task, not UI.

Silhouette rule. Platforms read as separate objects through contrasting silhouette. Dark platforms on light background, or light on dark — but not equal brightness. If platform blends with background, player misses jump — not from imprecise control, but from visual confusion.

Color code for hazards. Spikes — warm colors (red, orange). Acid — green. Lava — red-orange. Electric traps — yellow/blue. This is genre convention, violating without reason is mistake. Player knows these rules from other games.

Guide lines. Parallax layers, lighting, object placement create "sight lines" directing the player. Coin trail leads to bonus area. Brighter lighting at corridor end — goal is there. This is architectural solution, not UI element.

Location work structure

1. Reference and mood board — analogs from genre games + unique style elements
2. Blockout — rough geometry layout with colliders, gameplay test. No final art
3. Layer scheme — determining number and type of layers, parallax coefficients
4. Tileset design — tiles considering Rule Tile rules, PPU, sizes
5. Art pass — final graphics, lighting, atmospheric effects
6. Polish — animated elements (grass, water), particle effects, ambient sounds
7. QA — collider test, readability test on different screen resolutions

Cost determined by location size, detail level, and unique element count.