Expert audio engineer for interactive media: games, VR/AR, and mobile apps. Specializes in spatial audio, procedural sound generation, middleware integration, and UX sound design.
DECISION POINTS
1. Middleware Selection: Wwise vs FMOD
IF (budget < $10k AND indie game):
└─ Use free FMOD (up to $500k revenue)
IF (AAA production OR need extensive audio design tools):
└─ Use Wwise
└─ IF (team has dedicated audio programmers):
└─ Full Wwise SDK integration
└─ ELSE (programmers need simple API):
└─ Use Wwise Unity/Unreal plugins
IF (mobile-only OR web deployment):
└─ Consider lightweight alternatives
└─ Check platform restrictions (iOS/WebGL)
2. Spatial Audio Approach Selection
Decision Matrix Based on Context:
Sources > 20 AND VR with head tracking:
└─ Use Ambisonics (encode once, rotate cheaply)
Sources < 10 AND close/important sounds:
└─ Use full HRTF convolution per source
Mobile OR CPU budget tight:
└─ IF (stereo headphones expected):
└─ Simple binaural panning
└─ ELSE:
└─ Standard stereo panning + distance rolloff
Background/ambient sounds:
└─ Always use simple panning (save CPU for foreground)
3. Adaptive Music Strategy
IF (music needs to match gameplay intensity):
└─ Horizontal Re-orchestration:
└─ Layer 1: Basic rhythm/bass
└─ Layer 2: Add melody
└─ Layer 3: Add harmony/counterpoint
└─ Layer 4: Full orchestration
IF (different moods needed per area):
└─ Vertical Stems:
└─ Peaceful stem (woodwinds, strings)
└─ Tense stem (brass, percussion)
└─ Combat stem (full orchestra)
└─ Crossfade based on game state
IF (seamless transitions critical):
└─ Use musical bars as transition boundaries
└─ Pre-calculate next transition point
└─ Never cut mid-phrase
4. Footstep Implementation Choice
Memory budget > 5MB AND < 50 total surfaces:
└─ Use sample library approach
Memory budget tight OR > 100 surface variations:
└─ Procedural synthesis:
└─ Impact component (filtered noise burst)
└─ Surface texture (material-specific)
└─ Debris scattering (micro-impacts)
Performance critical (mobile):
└─ Pre-generate variations at load time
└─ Cache 10-20 variants per surface type
FAILURE MODES
1. HRTF Overload ("Everything Needs 3D")
Symptom: Frame drops when 20+ sounds play simultaneously
Detection: CPU profiler shows >50% time in HRTF convolution
Root Cause: Applying full HRTF to every sound source
Fix: Use HRTF only for 3-5 important sources; simple panning for background
: Set source importance hierarchy at design time
Prevention
2. Sample Memory Bloat ("Footstep Explosion")
Symptom: 500MB+ audio assets for simple character movement
Detection: 50+ footstep samples per character/surface combination
Root Cause: Artist creating samples for every possible variation
Fix: Implement procedural footstep synthesis with 4-5 base components
Prevention: Establish memory budgets early; use procedural for high-variation content
3. Mobile Session Chaos ("The Silent Treatment")
Symptom: App audio stops working after phone call/notification
Detection: Audio stops, never resumes; works fine on first launch
Root Cause: No interruption handling for iOS/Android audio sessions
Fix: Implement proper session management with interruption observers
Prevention: Test with incoming calls, music apps, Siri activation
4. UI Sound Assault ("Click Fatigue")
Symptom: Users disable sound after 10 minutes of interaction
Detection: Every button click at same volume as gameplay audio
Root Cause: UI sounds mixed at gameplay levels (-6dB instead of -20dB)
Fix: Reduce UI sounds to -18 to -24dB; use subtle, brief tones
Prevention: Follow platform audio guidelines; A/B test with real users
5. Real-time Processing Overload ("DSP Death Spiral")
Symptom: Audio stutters, pops, or cuts out during intense scenes
Detection: Audio callback exceeds allocated time budget (>10ms)
Root Cause: Too many real-time effects, unoptimized convolution
Fix: Use FFT-based convolution; limit concurrent DSP effects
Prevention: Profile on lowest-spec target device; set hard limits
WORKED EXAMPLES
Example 1: VR Footstep System (Procedural Approach)
Scenario: VR game needs infinite footstep variation across 15 surface types, tight memory budget (50MB total audio)
Novice: "Use realistic footstep recordings" → 150MB, repetition after 30 mins
Expert: "Synthesis sounds 90% as good, infinite variation, 50KB total"
Example 2: Mobile UI Sound Design (Session Handling)
Scenario: Meditation app needs subtle notification sounds that respect music apps and phone calls
Expert Decision Process:
Audio session category: .ambient with .mixWithOthers (don't interrupt Spotify)
Volume levels: -20dB for notifications, -24dB for UI feedback
Interruption handling: Pause during calls, resume after
Haptic coordination: Match audio transients to taptic feedback
Implementation:
classAppAudioManager {
funcsetupAudioSession() {
let session =AVAudioSession.sharedInstance()
try? session.setCategory(.ambient, mode: .default, options: [.mixWithOthers])
// Handle interruptions (phone calls, Siri)NotificationCenter.default.addObserver(
self, selector: #selector(handleInterruption),
name: AVAudioSession.interruptionNotification, object: nil
)
}
@objcfunchandleInterruption(notification: Notification) {
guardlet info = notification.userInfo,
let typeValue = info[AVAudioSessionInterruptionTypeKey] as?UInt,
let type =AVAudioSession.InterruptionType(rawValue: typeValue) else { return }
switch type {
case .began:
pauseAllAudio() // Phone call startedcase .ended:
iflet optionsValue = info[AVAudioSessionInterruptionOptionKey] as?UInt {
let options =AVAudioSession.InterruptionOptions(rawValue: optionsValue)
if options.contains(.shouldResume) {
resumeAudio() // Safe to resume
}
}
}
}
}
What Novice Misses: Assumes audio "just works"; ignores platform session requirements
Expert Insight: Mobile audio requires explicit session management; test with real interruptions
Example 3: Adaptive Music System (RTPC Setup)
Scenario: Action RPG needs music that scales with combat intensity (0 = exploration, 1 = boss fight)
Expert Decision Process:
Choose horizontal orchestration over vertical stems (smoother intensity scaling)
