Audio

Lumina plays sound through two components, a source that emits and a listener that hears. Every voice routes through a mix bus so groups of sounds can be balanced together, and one-off sounds can be fired straight from code.

Audio Source

Add an Audio Source component to an entity to make it emit a sound. Its transform drives distance attenuation, panning, and doppler.

Property	Does
Sound	The audio asset to play.
Bus	Mix group the voice routes through (SFX by default).
Volume	Volume multiplier (1.0 = full).
Pitch	Pitch multiplier (1.0 = original).
Spatialized	Off plays the sound flat, ignoring the transform.
Attenuation	Distance falloff, cone, and doppler settings. See below.
Occlusion	Muffling when geometry blocks the listener. See below.
Priority	Higher priority voices survive when the voice limit is reached.
Fade In Time	Seconds to ramp up from silence when playback starts.
Fade Out Time	Seconds to ramp down when the source is asked to fade out.
Looping	Restart automatically when it finishes.
Play On Ready	Start playing as soon as the entity is set up.
Cull Beyond Max Distance	Don’t start a voice the listener can’t hear anyway.

The component’s fields are settable at runtime, and it exposes a few methods.

C++

SAudioSourceComponent& Source = Context.Get<SAudioSourceComponent>(Entity);
Source.Volume = 0.5f;
Source.bLooping = true;
Source.Play();

C#

SAudioSourceComponent Source = Registry.Get<SAudioSourceComponent>(Entity);
Source.Volume = 0.5f;
Source.bLooping = true;
Source.Play();

The methods are identical in both languages.

Method	Effect
Play()	Start (or restart) the source.
Stop()	Cut the voice immediately.
FadeOut()	Ramp down over Fade Out Time, then stop.
SetPaused(bool)	Pause or resume without losing the playback position.
IsPlaying()	True while the mixer still holds a voice for this source.
GetPlaybackTime()	Current position in seconds.
SeekToTime(float)	Jump to a position in seconds.

Audio Listener

The Audio Listener component marks the “ears”. Sound is panned and attenuated relative to its world position, so put one on your camera or player. Without a listener, audio plays from the world origin.

Property	Does
Listener Index	Which of the four engine listener slots this component drives.
Apply Doppler	Feed the listener’s motion into the doppler calculation.

Split screen uses one listener component per view, each with its own index. Sounds are spatialized against whichever listener is closest.

Attenuation

The Attenuation struct on a source controls how a voice behaves in 3D.

Property	Does
Model	Inverse (default, physically correct), Linear, Exponential, or None.
Min Distance	Distance (m) the sound stays at full volume within.
Max Distance	Distance (m) the falloff curve bottoms out at.
Rolloff	Falloff steepness past Min Distance. Higher gets quiet sooner.
Min Gain / Max Gain	Floor and ceiling on the attenuated volume.
Cone Inner Angle	Degrees. Full volume inside this cone around the source’s forward axis.
Cone Outer Angle	Degrees. Volume ramps to Cone Outer Gain between the two angles.
Cone Outer Gain	Volume outside the outer cone.
Doppler Factor	Pitch shift from relative motion. 0 disables doppler for this voice.
Directional Factor	How much the listener’s facing attenuates the voice. 0 ignores facing.
Pan	Stereo pan applied after spatialization.
Positioning	Absolute (world space) or Relative (to the listener).

Leave the cone angles at 360 for an omnidirectional source. Narrow them for things that project sound in one direction, like a speaker or a megaphone.

Doppler needs velocity, which the engine derives automatically from how the source and the listener move each frame. A global Doppler Scale in the audio settings scales every voice at once.

Occlusion

Occlusion muffles a sound when level geometry sits between it and the listener. It’s off by default. Turn it on per source under Occlusion.

Property	Does
Enabled	Trace to the listener and muffle this source when blocked.
Low Pass Frequency	Cutoff (Hz) at full occlusion. Lower is more muffled.
Volume Attenuation	Volume multiplier at full occlusion.
Interp Time	Seconds to blend between occluded and clear. Keeps the filter from popping.
Trace Interval	Seconds between traces for this source.

The engine raycasts from the listener to the source, then blends the result in over Interp Time and applies it as a low-pass filter plus a volume drop. Sources that are never occluded pay nothing, the filter is only created the first time a source is actually blocked.

Which collision layers count as blocking, the per-tick trace budget, and a global on/off switch all live in the audio settings.

Distant or quiet sources don’t need frequent traces. Raise Trace Interval on ambience and background loops so the budget goes to sounds the player is paying attention to.

Voice limits

The mixer holds a limited number of simultaneous voices (128 by default, configurable up to 256). When they’re all in use, a new sound evicts the lowest priority voice that ranks below it, or is dropped if nothing does.

Two properties keep the budget healthy:

• Priority ranks a source against the rest. Give gunshots and dialogue a high value and ambience a low one.
• Cull Beyond Max Distance stops a source from taking a voice when the listener is outside its Max Distance. Looping sources that also have Play On Ready go further and release their voice when the listener walks away, taking a new one when they come back, so a level full of ambient loops costs nothing while you’re nowhere near it.

Procedural audio

A Procedural Audio component streams PCM samples you generate at runtime, for synths, engine sounds, or network voice. Set its sample rate and channel count, call Start(), then push interleaved float samples with QueueSamples(). It carries the same Bus and Attenuation properties as a regular source.

Next

• Playing sounds, firing and steering one-off voices from code.
• Mixing & Buses, mix groups, an options menu, and reverb.
• Audio Settings, project-wide defaults.