Skip to content

Audio and Sound Design

Introduction

Game audio is interactive: sound responds to player action and game state in real time, unlike linear film audio. It splits into three jobs — sound effects (feedback), music (emotion/pacing), and dialogue/VO (narrative) — all balanced by a runtime mix. This page covers the pipeline, adaptive music, spatialization, and the budget.

The audio pipeline

From source asset to the player’s speakers:

  1. Assets — recorded or synthesized samples, usually compressed (Vorbis/Opus) for size, kept uncompressed (PCM) for short latency-critical SFX.
  2. Events / cues — gameplay triggers an abstract event (“footstep”, “explosion”) rather than a specific file; the audio system picks the actual sample(s).
  3. Voices / channels — each playing sound occupies a voice; a finite voice budget is enforced by priority and culling.
  4. Buses / submixes — voices route through a mix graph (SFX bus, music bus, VO bus) for grouped volume, effects, and ducking.
  5. Master out — final limiter, output to the device at the platform’s sample rate.

SFX design principles

  • Variation — randomize pitch/volume and round-robin samples so repeated sounds (footsteps, gunfire) don’t fatigue the ear.
  • Layering — build one impactful sound from layers (a punch = whoosh + impact + low thud).
  • Feedback — every meaningful player action needs an audible confirmation; silence reads as a bug.
  • Readability — gameplay-critical sounds (enemy behind you, low health) must cut through the mix; cosmetic sounds yield.

Adaptive and interactive music

Music that responds to the game instead of looping flatly:

Horizontal re-sequencing

Arrange the track as segments and reorder/transition between them based on state (explore → tension → combat), respecting musical bar boundaries.

Vertical layering

Author stacked stems (drums, strings, lead) and fade layers in/out with intensity — same loop, rising energy.

Transitions & stingers

Quantize transitions to the beat; fire short stingers on events (victory, discovery) over the running bed.

State-driven

Game state (combat, menu, cutscene) drives the music system as inputs, not hardcoded track swaps.

Spatialization

Placing sound in 3D so the player can locate it:

  • Panning & attenuation — position in the stereo/surround field and roll off volume with distance.
  • 3D audio / HRTF — head-related transfer functions simulate how ears localize sound, for headphone immersion.
  • Occlusion & obstruction — muffle and low-pass sounds behind walls so geometry shapes what you hear.
  • Reverb zones — environment-driven reverb (cave vs hall vs open field) grounds sound in space.
  • Doppler — pitch shift on fast-moving sources (vehicles, projectiles).

The mix and budget

  • Voice budget — a hard cap on simultaneous sounds; priority and distance culling decide who plays when the cap is hit.
  • Ducking / sidechain — automatically dip music/SFX when dialogue plays so speech stays intelligible.
  • Dynamic range — leave headroom; a wall of max-volume sound is exhausting and hides important cues.
  • Memory & streaming — stream long music from disk, keep short latency-critical SFX resident in memory.

How KBVE applies these

  • Event Queue for audio — KBVE drains cross-system messages on a fixed cadence rather than firing direct calls; audio events fit that same Event Queue model — decouple when triggered from when played.
  • State-driven music — the GDD’s core-loop and game-state sections drive which audio state is active (exploration, combat, menus, cutscenes), matching the music-direction guidance in the core template.

Further reading