majiayu000 Audio Integration
How audio elements work together in a game: mixing, prioritization, spatial positioning, and dynamic systems.
Mix Architecture
Standard Bus Structure
[Master Bus]
┌────────────┼────────────┐
[Music] [SFX] [Voice]
│ │
┌──────┴──────┐ ┌──┴───┬─────┐
[Ambient] [Combat] [UI] [World] [Creatures]
Bus Purposes
| Bus | Contents | Processing |
|---|---|---|
| Master | Everything | Final limiter |
| Music | All music | Ducking input |
| SFX | All effects | Compression |
| Voice | Dialogue | De-esser, comp |
| UI | Interface | Dry, consistent |
→ See references/bus-templates.md for full configurations.
Mix Priorities
Not all sounds are equal:
| Priority | Category | Behavior |
|---|---|---|
| 1 (Critical) | Player damage, alerts | Always audible |
| 2 (High) | Player actions, dialogue | Clear and present |
| 3 (Medium) | Enemy sounds, music | Supporting |
| 4 (Low) | Ambient details | Background |
| 5 (Background) | Environmental loops | Atmospheric |
Voice stealing: When max voices reached, steal from lower priorities.
Ducking
Lower some sounds when others play.
Common Relationships
| Trigger | Target | Amount | Purpose |
|---|---|---|---|
| Voice | Music | -6 to -12dB | Hear dialogue |
| SFX | Music | -3 to -6dB | Hear actions |
| Critical | All | -6dB | Emergency alerts |
Parameters:
- Threshold: Level that triggers ducking
- Duck amount: How much to reduce (dB)
- Attack: How fast to duck (10-50ms)
- Release: How fast to recover (200-500ms)
→ See references/ducking-recipes.md for implementation patterns.
Spatial Audio
Distance Attenuation
| Sound Type | Min Dist | Max Dist | Curve |
|---|---|---|---|
| Voice | 1m | 15m | Linear |
| Footsteps | 0.5m | 20m | Inverse |
| Gunshots | 5m | 200m | Logarithmic |
| Ambient point | 2m | 50m | Linear |
| UI | N/A | N/A | 2D (no falloff) |
Occlusion
When sounds are blocked:
- Reduce volume (50% at full occlusion)
- Apply low-pass filter (20kHz → 500Hz)
- Muffled = blocked by walls
→ See references/spatial-audio.md for panning and occlusion code.
Reverb Zones
| Space | Decay | Early Reflections | Wet Level |
|---|---|---|---|
| Small room | 0.3-0.8s | High | 20-40% |
| Large hall | 1.5-3s | Medium | 30-50% |
| Cathedral | 4-8s | Low | 40-60% |
| Outdoor | 0.1-0.3s | Very low | 10-20% |
| Cave | 2-5s | High | 50-70% |
Per-sound reverb sends:
- Dialogue: Low (clarity)
- Footsteps: Match environment
- Music: Baked-in reverb
Dynamic Range
Compression by Context
| Context | Ratio | Attack | Release |
|---|---|---|---|
| Dialogue | 3:1 | 10ms | 200ms |
| SFX Bus | 2:1 | 1ms | 100ms |
| Music | 2:1 | 10ms | 300ms |
| Master | 4:1 | 5ms | 200ms |
Loudness Targets
| Platform | Target | True Peak |
|---|---|---|
| PC/Console | -14 LUFS | -1 dBTP |
| Mobile | -16 LUFS | -1 dBTP |
Audio States
Change audio based on game state:
| State | Music | SFX | Ambient | Notes |
|---|---|---|---|---|
| Menu | 100% | 50% | 0% | Focus on music |
| Exploration | 80% | 100% | 100% | Full soundscape |
| Combat | 100% | 110% | 50% | SFX priority |
| Cutscene | 100% | 80% | 60% | Dialogue priority |
| Paused | 30% | 0% | 0% | Minimal |
Transitions: Crossfade between states (500ms-2s).
→ See references/audio-states.md for state machine patterns.
Performance
Voice Management
- Max voices: 32-64 typical
- Priority-based stealing
- Distance-based LOD (reduce update rate for distant sounds)
Memory Management
| Load Type | Size | Use Case |
|---|---|---|
| Resident | < 500KB | Frequent sounds |
| Streaming | > 500KB | Music, long ambient |
| On-demand | Rare | One-time sounds |
Testing Checklist
- All sounds play without errors
- No clipping at master bus
- Priorities work (spawn 100 sounds)
- Ducking sounds natural
- Spatial audio pans correctly
- State transitions smooth
- Performance within budget
Related Components
sonic-style-languageskill - Style consistencyaudio-coherence-revieweragent - Validate integration/establish-sonic-identity- Set mix priorities