/ue-audio-system

name: ue-audio-system
description: "Use this skill when working with audio, sound, music, UAudioComponent, PlaySoundAtLocation, SoundCue, MetaSound, attenuation, submix, concurrency, SFX, or spatial audio in Unreal Engine. See references/audio-setup-patterns.md for music system and ambient soundscape architectures. For VFX audio synchronization, see ue-niagara-effects."
metadata:
  version: 1.0.0

UE Audio System

You are an expert in Unreal Engine's audio systems, covering UAudioComponent, sound asset types, spatial attenuation, concurrency management, submix routing, MetaSounds, and runtime audio analysis.

Context Check

Before implementing audio, read `.agents/ue-project-context.md` for:

• **Audio plugins** enabled (Resonance Audio, Steam Audio, Wwise, FMOD, MetaSound plugin version)
• **Target platforms** — mobile has strict voice limits; consoles differ from PC
• **Dedicated server** flag — audio must be skipped server-side or it will crash/log errors
• **VR flag** — VR projects require binaural spatialization settings

Information Gathering

Ask about:

1. One-shot SFX, looping ambient, music, UI feedback, or dialogue? 2. Spatialized (follows actor) or global 2D? 3. Concurrency concern (gunshots, footsteps, explosions)? 4. Runtime control needed (fade, pause, parameter changes)? 5. MetaSound procedural or pre-authored SoundCue/SoundWave?

---

Sound Asset Hierarchy

USoundBase                    // abstract base (SoundBase.h)
  ├── USoundWave              // raw PCM/compressed audio asset
  ├── USoundCue               // node-graph: random, modulator, mixer, attenuator nodes
  └── UMetaSoundSource        // procedural audio graph (MetaSound plugin)

**USoundWave** — Import .wav/.ogg/.flac. Set `SoundClassObject` and `AttenuationSettings` on asset.

**USoundCue** — Node graph combining multiple waves. Key nodes: `USoundNodeRandom`, `USoundNodeModulator`, `USoundNodeMixer`, `USoundNodeAttenuation`, `USoundNodeLooping`, `USoundNodeDelay`, `USoundNodeDistanceCrossFade`.

**UMetaSoundSource** — Procedural audio graph. Declare typed inputs (float, bool, int32, trigger). Set parameters at runtime via `UAudioComponent::SetFloatParameter`, `SetBoolParameter`, `SetIntParameter`.

Streaming Long Audio

For music and ambient tracks exceeding ~30 seconds, set `USoundWave::LoadingBehavior`: `ESoundWaveLoadingBehavior::ForceInline` for short SFX, `RetainOnLoad` for music loaded at level start. Long files should use `LoadOnDemand` to avoid loading the full waveform into memory. In the editor: SoundWave asset → Details → Loading → Loading Behavior.

---

Playing Sounds from C++

Fire-and-Forget

#include "Kismet/GameplayStatics.h"

// 2D — not spatialized (UI, music)
UGameplayStatics::PlaySound2D(
    this, ImpactSound, 1.0f /*Vol*/, 1.0f /*Pitch*/, 0.0f /*StartTime*/,
    ConcurrencySettings, OwningActor
);

// 3D — spatialized, requires AttenuationSettings on the sound asset
UGameplayStatics::PlaySoundAtLocation(
    this, GunShotSound, GetActorLocation(), FRotator::ZeroRotator,
    1.0f, 1.0f, 0.0f,
    AttenuationOverride,    // USoundAttenuation* (nullptr = use asset default)
    ConcurrencyOverride,    // USoundConcurrency* (nullptr = use asset default)
    this                    // OwningActor for per-owner concurrency
);

Spawn with Handle

// Returns UAudioComponent* — auto-destroyed when sound finishes if bAutoDestroy=true
UAudioComponent* Comp = UGameplayStatics::SpawnSoundAtLocation(
    this, ExplosionSound, Location, FRotator::ZeroRotator,
    1.0f, 1.0f, 0.0f, AttenuationSettings, nullptr, /*bAutoDestroy=*/true
);

// Attach to a moving component (vehicle engine)
UAudioComponent* EngineAudio = UGameplayStatics::SpawnSoundAttached(
    EngineLoopSound, GetMesh(), NAME_None,
    FVector::ZeroVector, FRotator::ZeroRotator,
    EAttachLocation::SnapToTargetIncludingScale,
    /*bStopWhenAttachedToDestroyed=*/true,
    1.0f, 1.0f, 0.0f, AttenuationSettings, nullptr,
    /*bAutoDestroy=*/false   // keep alive for looping
);

UAudioComponent as Permanent Actor Component

// In constructor:
AudioComponent = CreateDefaultSubobject<UAudioComponent>(TEXT("AudioComponent"));
AudioComponent->SetupAttachment(RootComponent);
AudioComponent->bAutoActivate = false;
AudioComponent->bStopWhenOwnerDestroyed = true;

Playback Control

AudioComponent->SetSound(EngineLoopSound);
AudioComponent->Play(/*StartTime=*/0.0f);
AudioComponent->Stop();
AudioComponent->SetPaused(true);
AudioComponent->FadeIn(0.5f, 1.0f, 0.0f, EAudioFaderCurve::Linear);
AudioComponent->FadeOut(1.0f, 0.0f, EAudioFaderCurve::Linear);
AudioComponent->SetVolumeMultiplier(0.5f);
AudioComponent->SetPitchMultiplier(1.2f);

// Query play state (EAudioComponentPlayState: Playing, Stopped, Paused, FadingIn, FadingOut)
EAudioComponentPlayState State = AudioComponent->GetPlayState();

Delegates (AudioComponent.h)

// Declared in AudioComponent.h:
// DECLARE_DYNAMIC_MULTICAST_DELEGATE(FOnAudioFinished)
// DECLARE_DYNAMIC_MULTICAST_DELEGATE_TwoParams(FOnAudioPlaybackPercent, const USoundWave*, PlayingSoundWave, const float, PlaybackPercent)
// DECLARE_DYNAMIC_MULTICAST_DELEGATE_OneParam(FOnAudioPlayStateChanged, EAudioComponentPlayState, PlayState)

AudioComponent->OnAudioFinished.AddDynamic(this, &AMyActor::OnSoundFinished);
AudioComponent->OnAudioPlaybackPercent.AddDynamic(this, &AMyActor::OnPlaybackPercent);
AudioComponent->OnAudioPlayStateChanged.AddDynamic(this, &AMyActor::OnPlayStateChanged);

// Native (non-UObject) binding — no GC overhead:
// DECLARE_MULTICAST_DELEGATE_OneParam(FOnAudioFinishedNative, UAudioComponent*)
// DECLARE_MULTICAST_DELEGATE_ThreeParams(FOnAudioPlaybackPercentNative, const UAudioComponent*, const USoundWave*, const float)
AudioComponent->OnAudioFinishedNative.AddUObject(this, &AMyActor::OnSoundFinishedNative);
AudioComponent->OnAudioPlaybackPercentNative.AddUObject(this, &AMyActor::OnPlaybackPercentNative);
`