Bleeps, Blops, Clicks, and Pops

Sound Design in Games

Adaptive Audio

Learning outcomes

  • Map a practical game-audio workflow from asset list to runtime behavior
  • Differentiate interactive versus adaptive audio using observable criteria
  • Classify layers and propose mix and accessibility strategies

From Foley to parameters

  • Theatrical sound effects and Foley established timing and gesture
  • Digital libraries and DAWs scaled content creation and layering
  • Middleware and parameters enabled systems that respond at runtime

Interactive versus adaptive

  • Interactive: audio is triggered directly by player input or discrete events
  • Adaptive: audio changes in response to game state or parameter values
  • Common mappings
    • Input → one-shot SFX
    • State → music section or snapshot
    • Parameter → continuous change of pitch, filter, send level

Diegetic and non-diegetic

  • Diegetic: exists in the world; characters could hear it
  • Non-diegetic: player-facing guidance or score
  • Ambiguous cases to debate
    • In-world radio that crossfades to score
    • UI sounds that are framed as in-world devices

Layered sound taxonomy

LayerPrimary goalPitfallsMix tools and policies
AmbienceSense of place and timeMasking dialogueEQ dips, level, bus ducking
World/FoleyMaterial identity, affordancesRepetition, machine-gun effectVariations, random start, filters
UI/UXFeedback and guidanceLoudness spikes, overuseLoudness targets, short decay
Dialogue/VOIntelligibilityFighting music/ambienceSidechain, clarity EQ
MusicNarrative pacing and emotionLoop fatigue, clashesStingers, states, blends

Platform realities and mix strategy

  • Constrained platforms: pre-mix, consolidate layers, bake variation into assets
  • Capable platforms: runtime mixing, snapshots, sends, real-time effects
  • Decide what you pre-bake versus what the engine controls

Choosing file formats

  • There is no single codec that fits every audio use case; each has strengths for different scenarios.
  • Opus is best for compressing voice and dialog, combining high quality with small file sizes.
  • WAV/PCM offers the fastest decoding and no loss of quality but results in very large files.
  • Vorbis is a strong, widely compatible alternative when Opus isn’t available.
  • Hardware or software decoding methods can affect performance, so codec choice may depend on platform support.

Source

Start with an asset list

  • Name, category, trigger, diegetic flag
  • Playback (one-shot or loop), variations, randomization
  • Parameters, ducking, priority, spatialization
  • File format, SR/bit depth, loudness target, owner, status
IDAsset NameCategoryTrigger/EventDiegeticPlayback
01footstep_grass_lightFoleyplayer_stepYesOne-shot
02door_metal_openFoleyinteract_openYesOne-shot
03ui_inventory_openUI/UXui_openNoOne-shot
IDVariationsRandomizationParametersDuckingFormat
018start ±20%, pitch ±3 stplayer_speedUI bus −4 dB, 50/200OGG
023start ±10%, vol ±2 dBnonemusic bus −3 dBWAV
032nonenonenoneOGG

Mini-lab: footsteps that adapt

Goal: Design a small assets list for footsteps that adapt to player_speed and surface.

Template

Steps

  • Groups of 3–4 choose one surface: grass, metal, water, stone
  • In the class Google Sheet (tab for your surface), fill 5 rows using: Asset Name · Category · Trigger/Event · Diegetic · Playback · Variations · Randomization · Parameters/RTPCs · Ducking · Priority · File Format
  • In Notes: define mappings (e.g., player_speed → pitch ±3 st, vol ±3 dB; crouch → high-cut)
  • Prepare a 1-minute rationale: how you avoid repetition and masking
  • Deliverable: Five rows completed + 1-minute talk-through

Roles and collaboration thread

  • Audio Director or Lead: loudness targets and scope guardrails
  • Sound Designer: variation design, naming, and mix intent
  • Audio Programmer: parameter plumbing, snapshots, performance budgets

Appendix: optional early tech clip

Mike Patton and the Intonarumori