Audio¶

UE 5.7's audio stack has three eras of tech:

Era	What
Legacy	`USoundCue` (still works, still useful)
Modern	MetaSounds (procedural DSP)
Middleware	Wwise / FMOD plugins (industry standard for AAA)

Pages¶

MetaSounds — modern procedural audio
Sound Cues — legacy graph
Sound Classes — bus / category routing
Audio Volumes — region-based reverb / occlusion
Quartz — sample-accurate scheduling

Asset taxonomy¶

Asset	Role
`USoundWave`	Audio file
`USoundCue`	Legacy DSP-light graph
`UMetaSoundSource`	Modern procedural source
`UMetaSoundPatch`	Reusable MetaSound subgraph
`USoundClass`	Group of sources with shared properties
`USoundSubmix`	Output routing bus
`USoundAttenuation`	3D distance / cone settings
`UReverbEffect`	Reverb preset
`USoundMix`	Mix snapshot (volume per class)
`UDialogueWave`	Localized dialogue
`USoundConcurrency`	Voice-count limits

Signal flow¶

graph LR
  SC[SoundCue / MetaSound] --> SS[SoundClass]
  SS --> SX[SoundSubmix]
  SX --> M[Master Submix]
  M --> O[Audio Out]
  V[AudioVolume] -.->|reverb / atten override| SC

Choosing source type¶

Need	Use
Static audio with randomization	SoundCue or MetaSound
Procedural sound (engine, footsteps with timing)	MetaSound
Cross-fades, transitions, complex mixes	MetaSound
AAA audio production with middleware-grade authoring	Wwise / FMOD
Music with sample-accurate beats	MetaSound + Quartz

Performance¶

Concurrency settings cap simultaneous voices per source.
Profile with stat audio, au.audiothread.usagetreshold.
MetaSounds are evaluated per-source on the audio thread — heavy graphs add CPU.