audio-design

What this skill does

# Audio Design

Audio is the fastest feedback channel in a game. Players process sound faster than
visual information — audio tells them what is happening before they see it, behind
them, and outside their field of view. A game where players mute the audio and lose
nothing has failed at audio design.

This skill covers audio as a **design system**: deciding what sounds exist, when they
play, how they relate to gameplay state, and how they serve the player's understanding
of the game world. Production concerns (recording, mastering, compression formats) are
out of scope.

---

## When to Activate

- Designing sound effects for player actions, enemies, or environment
- Building adaptive music systems that respond to gameplay
- Planning spatial audio for directional awareness and immersion
- Creating ambient soundscapes that establish atmosphere
- Diagnosing why audio feels disconnected from gameplay
- Noticing that players routinely play on mute
- Adding audio accessibility features
- Evaluating audio feedback coverage across game systems
- Balancing the audio mix and priority system

---

## Core Principle: Audio as Information System

Sound is not decoration — it is an information channel that operates independently from
the visual field. Design audio with the same rigor applied to any other game system.

**What audio communicates that visuals cannot:**

| Advantage | Example |
|---|---|
| 360-degree awareness | Footsteps behind the player |
| Off-screen events | Explosion around a corner |
| Anticipation/warning | Audio wind-up before an attack lands |
| Emotional context | Minor-key shift signaling danger |
| Confirmation | Satisfying click confirming a successful action |
| Pacing | Tempo change signaling urgency |

**The mute test:** If a player can play the game on mute with no loss of information or
enjoyment, the audio design is not carrying its weight. Every audio element should either
convey information, create emotion, or both. Purely decorative audio that does neither is
wasted bandwidth in the mix.

---

## The Audio Stack

Audio architecture uses independent layers, each serving a distinct purpose. Every layer
must be independently controllable — players need per-layer volume sliders, not just a
master volume knob.

### Layer 1: Music

**Purpose:** Emotional context, tension management, pacing.

Music tells the player how to *feel* about what is happening. It operates on slower
timescales than other audio — setting mood across minutes rather than reacting frame by
frame.

- Establishes emotional tone for areas, encounters, and story beats
- Manages tension arcs (build, sustain, release)
- Creates anticipation for events that have not happened yet
- Signals safety or danger before the player has evidence

### Layer 2: Ambience

**Purpose:** Spatial context, atmosphere, world presence.

Ambience is the audio foundation that makes a space feel real. Without it, environments
feel sterile regardless of visual quality.

- Area-specific ambient beds (forest, cave, city, dungeon)
- Time-of-day variation (dawn chorus, night insects)
- Weather audio (rain, wind, thunder)
- Incidental environmental sounds (distant bird, creaking wood)

### Layer 3: Sound Effects

**Purpose:** Game state feedback, player action confirmation.

Sound effects are the primary feedback channel for moment-to-moment gameplay. They confirm
actions, communicate results, and signal state changes.

- Player action sounds (attack, jump, dash, interact)
- Impact and result sounds (hit, miss, block, break)
- Enemy feedback (telegraph, attack, damage, death)
- Environmental interaction (doors, switches, pickups)

### Layer 4: UI Audio

**Purpose:** Interface feedback, notifications, system communication.

UI audio confirms interface actions and draws attention to important system events. It
should be clean, consistent, and never fatiguing.

- Menu navigation (hover, select, back, error)
- HUD notifications (low health, buff applied, cooldown ready)
- Achievement and progression sounds
- Inventory and crafting feedback

### Layer 5: Voice

**Purpose:** Dialogue, barks, callouts, narrative delivery.

Voice is the most attention-demanding audio layer. When voice plays, everything else
should defer.

- Story dialogue and cutscene speech
- Companion/NPC barks (contextual one-liners)
- Combat callouts (enemy spotted, low ammo, revive request)
- Tutorial and hint voiceover

---

## Sound Effect Design

Every significant player action needs an audio response. Missing audio feedback creates
a disconnect that players feel even if they cannot articulate it.

### Feedback Principles

**Match visual weight.** A massive explosion with a thin pop sound creates cognitive
dissonance. A small UI toggle with a booming thud feels wrong. Audio and visual weight
must be proportional.

**Immediate onset.** Audio feedback should begin within 1–2 frames of the corresponding
visual onset. The perception-action cycle depends on audio and visual alignment — see
**game-feel** for timing reference tables. Late audio feels sluggish; early audio feels
disconnected.

**Priority ordering:**

| Priority | Category | Rationale |
|---|---|---|
| 1 | Player actions | The player must hear confirmation of their own inputs |
| 2 | Enemy feedback | Threats need clear audio communication |
| 3 | Environment | World interaction supports immersion |
| 4 | Ambience | Atmosphere fills gaps but yields to gameplay |

### Variation

Any sound that plays more than twice within a 10-second window needs at minimum 3
variants. Repetition fatigue is one of the fastest ways to make audio annoying.

**Variation techniques:**

- **Pitch randomization:** ±5–15% pitch shift on each play (cheap, effective)
- **Alternate clips:** 3–5 recorded or synthesized variants, round-robin or random
- **Layered randomization:** Randomize individual layers of a composite sound
- **Velocity scaling:** Vary volume and EQ based on action intensity

### Layering Complex Sounds

Complex actions deserve composite audio built from multiple layers:

```
Sword hit = impact layer + material layer + environment layer
         = [metal clang]  + [flesh thud]   + [room reverb tail]
```

Each layer can be swapped independently — the same impact with a different material
layer communicates hitting stone vs. wood vs. armor.

---

## Adaptive Music System

Static music that ignores gameplay state is a missed opportunity. Music should respond
to what the player is experiencing.

### Music States

Define discrete music states tied to gameplay context:

| State | Character | Typical Triggers |
|---|---|---|
| Exploration | Calm, open, melodic | No threats, traversal, discovery |
| Tension | Building, uneasy, rhythmic | Threat proximity, low resources, time pressure |
| Combat | Intense, driving, percussive | Active combat engagement |
| Resolution | Release, exhale, thinning layers | Combat end, threat eliminated |
| Victory | Triumphant, ascending, major key | Boss defeat, objective complete |
| Defeat | Somber, descending, sparse | Player death, mission failure |

### Transition Techniques

| Technique | Description | Best For |
|---|---|---|
| Crossfade | Blend out current, blend in next | Gradual mood shifts |
| Stinger-to-track | Short musical hit bridges to new track | Sudden state changes |
| Horizontal re-sequencing | Same tempo, swap musical layers | Intensity changes within a state |
| Vertical layering | Add/remove instrument layers | Smooth intensity scaling |

### Intensity Layers

The most flexible adaptive music approach uses vertical layering:

- **Base layer** always plays (establishes key, tempo, harmonic foundation)
- **Layer 2** adds rhythmic drive (percussion, bass patterns)
- **Layer 3** adds melodic tension (strings, synth leads)
- **Layer 4** adds full intensity (full orchestra, heavy percussion)

Layers add or remove based on a gameplay intensity value (0.0–1.0). This produces
smooth, musically coherent transitions because all layers share tempo