SoundTrace SDK for Unreal Engine

The SoundTrace Unreal Engine SDK is a real-time spatial audio plugin connecting Unreal audio sources, listeners, and meshes to STCoreV2. Configure acoustic paths, materials, HRTF, and GPU processing through project-wide settings and Actor Components.

Requirements and platforms

Item	Requirements / support
Unreal Engine	5.6 or newer
Supported platforms	Windows x64, macOS, Linux, Android, iOS

Installation

1. Close Unreal Editor and the target project.
2. Copy Plugins/SoundTracing from the SDK into the target project's Plugins/ directory.
3. Open the .uproject and enable the SoundTracing plugin.
4. Complete the C++ module build and restart the Editor.
5. Review the global settings under Project Settings > Plugins > SoundTracing.

YourProject/
└─ Plugins/
   └─ SoundTracing/
      ├─ Content/
      ├─ Source/
      └─ ThirdParty/

Keep Content/STData and ThirdParty/STCoreV2 when distributing the plugin folder. Material presets, custom HRTF data, and the native runtime use these paths.

Unreal Editor settings

1. Open the target platform's Audio settings under Edit > Project Settings > Platforms.
2. Set Spatialization Plugin to SoundTracing.
3. Set Source Data Override Plugin to SoundTracing.
4. Restart the editor.
5. Open Project Settings > Plugins > SoundTracing and review the global settings.

For Windows projects, start with the values below and adjust them to suit the audio workload.

Setting	Starting value
Audio Sample Rate	48000 Hz
Callback Buffer Frame Size	1024
Buffers To Enqueue	2

Audio asset import settings — mono

Prepare audio sources for SoundTrace spatialization as mono (one channel).

1. Export the original audio as a mono PCM WAV file from an audio editor.
2. Use Import in the Content Browser. Unreal creates a Sound Wave asset.
3. Confirm that the Sound Wave has 1 channel.
4. Assign the asset to the Audio Component's Sound field.

Mono describes the input source channels. SoundTrace renders headphone directionality relative to the listener. See Unreal audio import documentation for the basic import workflow.

Quick start

1. Start with Quality Preset = Fast in Project Settings > Plugins > SoundTracing.
2. Create a source settings asset from Sounds > SoundTracing > SoundTracing Audio Spatialization Settings in the Content Browser.
3. Create a Sound Attenuation asset and enable Enable Spatialization. Set Spatialization Method to Plugin-Spatialized and add the source settings asset to the Spatialization Plugin Settings array.
4. Assign a mono Sound Wave to the Audio Component. Enable Allow Spatialization, disable Override Attenuation, and assign the Sound Attenuation asset to Attenuation Settings.
5. Add SoundTracingObjectComponent as an immediate child of the StaticMeshComponent or SkinnedMeshComponent used for acoustic geometry.
6. Run Auto Set Materials on the object and adjust material slots as needed.
7. Start PIE and play the Audio Component. Check runtime readiness and path counts with SoundTracing.Status.
8. Add SoundTracingPathVisualizerComponent to an Actor to display paths.

Manage global configuration in SoundTracing Plugin Settings and query runtime state through SoundTracingSubsystem. Without a Listener Component, the plugin uses the Unreal audio listener transform and the project's default listener settings.

Component overview

Component	Unreal Engine SDK	Role
SoundTracing Plugin Settings	SoundTracingSettings, SoundTracingSubsystem	Project configuration and runtime state queries and control
Listener	SoundTracingListenerComponent	Listener quality, output settings, and transform overrides
Source	SoundTracingSourceSettings + Audio Component	Per-source emission, acoustic paths, and attenuation
Sound Object	SoundTracingObjectComponent	Registers meshes, BVH, and material slots
Acoustic Materials	SoundTracingMaterialPresetLibrary	Material presets and per-band reflection, absorption, and transmission
Sound Path Visualizer	SoundTracingPathVisualizerComponent	Displays acoustic paths with Niagara

SoundTracing Plugin Settings

Detail Properties

SoundTracingSettings manages global settings; SoundTracingSubsystem controls the runtime. Edit global configuration in Project Settings > Plugins > SoundTracing. Tables show SDK defaults; the images show example configurations.

Field	Default	Description
Propagation Thread Count	-1	-1..64. -1 configures threads automatically from logical core count; 0 and 1 use a single thread. Restart after changing.
Use GPU Backend	Disabled	Requests GPU processing, with CPU fallback if initialization fails. Restart after changing.
Path Cache Size	256	0..1024. Path cache shared by active sources. 0 disables caching.
Propagation Interval (ms)	0	0..500. Minimum interval between propagation requests. 0 requests on each game tick; requests coalesce while processing is in progress.
Quality Preset, Listener Rays, HRTF, Render Quality	Fast	Default listener profile. Each field is described in the listener Detail Properties table below.
Source Ray Resolution Cap	0	0..32. Global cap for per-source reverb ray resolution. 0 adds no limit.
Source Ray Depth Cap	0	0..16. Global cap for per-source reverb ray depth. 0 adds no limit.
Per-path Strength	each 1.0	Attenuation strengths for Direct, Reflection, Diffraction, Reverb, and Transmission, in 0.5..1.5. Applies to sources using project attenuation. Higher values attenuate faster at the same distance.
Material Preset Library	Unassigned	Uses the bundled library when unassigned. Restart after selecting another library.
Paths, Air Absorption	Enabled	Path families and air absorption for the default listener.

Public methods

In Blueprint, use Get World Subsystem to obtain SoundTracingSubsystem. In C++, use USoundTracingSubsystem::Get(WorldContextObject).

Method	Return value / behavior
Get(const UObject* WorldContextObject)	Static C++ method returning the subsystem for the specified World.
IsNativeRuntimeReady()	bool. Whether the native library, scene, and listener are ready.
IsGpuPropagationActive()	bool. Whether GPU propagation is actually active.
GetGpuBackendStatus()	FString: GPU active, CPU, or CPU fallback (...) with a reason. Returns Unavailable if the runtime cannot be accessed.
GetLastValidPathCount()	int32. Valid path count from the most recently completed propagation frame.
GetLastNativeError()	FString. Most recent native error; empty if the latest frame succeeded.
GetNativeVersion()	FString. Native library version in major.minor.revision format.
GetRegisteredObjectCount()	int32. Number of objects registered in the current World.
GetActiveListenerSettings()	FSoundTracingListenerSettings. Listener settings currently applied.
ResetMotionState()	Resets listener and source motion history after teleporting or changing levels.
RequestPropagationFrame()	Requests propagation outside the regular update cadence.

Public properties

These are the C++ configuration members of USoundTracingSettings. Read project settings with GetDefault<USoundTracingSettings>().

Property	Type	Description
PropagationThreadCount	int32	Configured propagation thread count.
bEnableGpuPropagation	bool	GPU request flag. Check actual activation with IsGpuPropagationActive().
PathCacheSize	int32	Path cache size.
PropagationIntervalMs	int32	Minimum propagation request interval in milliseconds.
DefaultListenerSettings	FSoundTracingListenerSettings	Project default listener profile.
SourceRayResolutionCap, SourceRayDepthCap	int32	Global resolution and depth caps for source rays.
DefaultSourceAttenuationStrengths	FSoundTracingAttenuationStrengths	Default distance attenuation strength for each path family.
MaterialPresetLibrary	TSoftObjectPtr<USoundTracingMaterialPresetLibrary>	Material library to register at startup.

Listener

Add SoundTracingListenerComponent to a Pawn or Camera to override the project's listener profile per level, or to use this component's transform instead of the Unreal audio-device listener.

Detail Properties

Project Settings and Listener Component share the listener profile fields. The image above shows the project's default profile. The component also exposes the override options below.

Field	Default	Behavior
Override Project Listener Settings	On	Applies Listener Settings instead of Project Settings
Listener Settings	Fast	Quality, HRTF, output mode, path, and air-absorption settings
Drive Listener Transform	Off	Sends this component's world transform and velocity to the native listener

Only one Listener Component drives the active listener in a World. If another component begins play, it replaces the previous component and writes a warning to the Output Log.

Listener profile field	Default	Description
Quality Preset	Fast	Selects Custom, Fast, Middle, or Quality.
Ray Resolution, Ray Depth	16, 4	1..32 and 1..16, respectively. Path tracing resolution and depth, editable in Custom.
Output Mode	Headphones	Selects headphone or speaker output.
Hrtf Mode	HRIR Interpolated	See the HRTF table below.
Custom HRTF Relative Path	Empty string	Custom HRTF path relative to plugin Content. Empty uses the embedded table.
HRTF Path Budget	1	1..32. Number of top paths receiving directional HRTF processing.
Diffuse Enabled, Diffuse Quality	Disabled, Low	Controls early scattered sound and its quality. Low/Medium/High retains up to 128/512/1024 scattered paths.
Delay Interpolation	Linear	Selects Linear, Cubic Lagrange, or Lagrange 6 delay interpolation.
Early Path Budget	128	0..4096. Early indirect paths receiving full moving-delay processing. 0 removes the limit.
Render Band Tier	Merged4	Selects the render band count with Merged4 or Full8.
Per-path Enable … Path	All enabled	Enables or disables Direct, Reflection, Diffraction, Reverb, and Transmission paths individually.
Air Absorption Enabled	Enabled	Applies attenuation from air absorption.
Temperature Celsius	20	-40..60 °C. Air temperature.
Relative Humidity Percent	50	0..100%. Relative humidity.
Pressure Pa	101325	50000..120000 Pa. Air pressure.

Quality Preset

Fast, Middle, and Quality apply the ray and advanced render-quality values together. Select Custom to edit them manually.

Preset	Ray Resolution	Ray Depth	HRTF Path Budget	Diffuse	Delay Interpolation	Early Path Budget
Custom	Stored value	Stored value	Stored value	Stored value	Stored value	Stored value
Fast	16	4	1	Off / Low	Linear	128
Middle	24	8	1	On / Medium	Cubic Lagrange	128
Quality	32	12	1	On / High	Lagrange 6	128

HRTF Path Budget is the number of highest-priority paths that receive a full directional HRTF voice. Every built-in preset keeps it at 1 for audio-thread headroom. Raising it in Custom mode can cause dropouts in scenes with many sources.

HRTF and output mode

HRTF mode	Required asset	Description
Band8	None	Lightweight 8-band magnitude and ITD mode
HRIR	STCoreV2 embedded table	Uses the nearest measured HRIR direction
HRIR Interpolated	STCoreV2 embedded table	Interpolates measured HRIRs by direction; the default

Set Custom HRTF Relative Path to an MPI1, SAH1, or BPH1 table path relative to plugin Content. Leave it empty to use the table embedded in STCoreV2.

Output Mode	Description
Headphones	Binaural HRTF output
Speaker	Internal Ambisonic stereo decode

The recommended advanced Render Band Tier is Merged4; Full8 increases per-path audio-thread band processing. Air Absorption defaults to 20 °C, 50% relative humidity, and 101325 Pa for ISO 9613-1 attenuation.

Public methods

Method	Behavior
ApplyListenerSettings()	Reapplies current Detail Properties values to the native listener
SetQualityPreset(Preset)	Changes and immediately applies the quality preset
SetHrtfMode(Mode)	Changes and immediately applies the HRTF mode
SetOutputMode(Mode)	Switches Headphones/Speaker and applies it
ResetMotionState()	Clears listener/source velocity history after a teleport or respawn
GetListenerSettings()	Returns the component's current listener settings
static GetActiveListener(const UWorld* World)	Returns the active Listener Component for the World in C++.

Public properties

Property	Type / access	Description
bOverrideProjectListenerSettings	bool / read/write	Whether to use this component's listener profile.
ListenerSettings	FSoundTracingListenerSettings / read/write	Component quality, output, and path settings. Call ApplyListenerSettings() after modifying them directly at runtime.
bDriveListenerTransform	bool / read/write	Whether the component's position and orientation drive the listener.

Source

Create this asset from Sounds > SoundTracing > SoundTracing Audio Spatialization Settings in the Content Browser. Share one asset between sources with the same role instead of duplicating it per Audio Component.

Detail Properties

Field	Default	Range/behavior
Intensity	1.0	0..10. Linear emission gain multiplied into every path
Gain Boost Db	0 dB	-24..24 dB. Additional gain above Intensity
Reverb Send Db	0 dB	-24..24 dB. Late-reverb send
Reflection Send Db	0 dB	-24..24 dB. Early-reflection send
Ray Preset	Custom	Custom, Fast (8×8, depth 4), Middle (16×16, depth 4), Quality (24×24, depth 4). Anything but Custom overwrites the two values below. This is the Unity SoundTraceSource Reverb Ray Resolution and applies to every source sharing the asset
Ray Resolution	24	0..32. 0 inherits the Listener grid; otherwise uses an N × N source reverb grid
Ray Depth	4	0..16. 0 inherits Listener depth
Per-path … Path Enabled	All enabled	Enables or disables Direct, Reflection, Diffraction, Reverb, and Transmission per source.
Override Project Attenuation Strengths	On	When off, uses the Project Settings attenuation values
Per-path Strength	1.0	0.5..1.5. Higher values attenuate faster at the same distance
Max Delay Seconds	1.0 s	0.01..5 s. Maximum propagation delay retained by the renderer
Path Fade Time Seconds	0.066 s	0.001..0.5 s. Fade time for path entry/removal
Path Hold Time Seconds	0.120 s	0..1 s. Holds a missing non-direct path before fading. 0 disables hold
Max Delay Rate	0.1	0.001..0.999. Maximum delay change per sample
Bypass	Off	Passes input through without SoundTrace spatial rendering

Connect to an Audio Component

Assign the source settings asset to Sound Attenuation, then assign that Attenuation asset to the Audio Component. Sources of the same kind can share both assets.

Enable Enable Spatialization and select plugin spatialization. Add the SoundTracing Audio Spatialization Settings asset to the Spatialization Plugin Settings array.

Enable Allow Spatialization and disable Override Attenuation on the Audio Component. Assign the previously created Sound Attenuation to Attenuation Settings.

Public methods

These are C++ methods on USoundTracingSourceSettings. Use the Audio Component's Play() and Stop() to control playback.

Method	Return value / behavior
GetEffectiveAttenuationStrengths()	FSoundTracingAttenuationStrengths. Returns source or project strengths according to the override setting.
ApplyRayPreset()	Updates RayResolution and RayDepth from RayPreset. Leaves values unchanged in Custom.

Sound Object

SoundTracingObjectComponent registers its immediate parent StaticMeshComponent or SkinnedMeshComponent as acoustic geometry. Add it directly below the target mesh component, not at an arbitrary place in the Actor hierarchy.

Detail Properties

Field	Default	Description
Auto Register Native Object	On	Registers the native object at Begin Play
Sync Transform On Tick	On	Updates the native transform only when the parent changes
Auto Sync Materials	On	Refreshes slots when parent render materials change
Update Type	Static	Geometry update policy
BVH Type	LBVH SIMD8	Native acceleration-structure builder
BVH Max Depth	12	1..32
Primitives Per Leaf	16	1..128
Sync Skinned Vertices On Tick	Off	Uploads the current skeletal pose each tick in Refit mode
Sound Material Slots	Generated	Maps render-material slots to SoundTrace presets
Visualize BVH	Off	Draws BVH lines with the Editor component visualizer
BVH Visualization Color	Cyan	BVH display color.
Native Object Id, Native Mesh Id	-1	Registered native IDs. Read-only; -1 when unregistered.

A Static Mesh uses Forced LOD when set, otherwise LOD 0. A Skinned Mesh also prefers Forced LOD and uploads the current pose vertices.

Geometry and BVH

BVH Type	Refit	GPU backend	Description
HKDTree	Supported	No	KD-split traversal; switches to BVH-style fallback after refit
LBVH	Supported	No	Scalar Morton LBVH
LBVH SIMD4	Supported	Yes	4-wide leaf intersection
LBVH SIMD8	Supported	Yes	Current default
LBVH SIMD16	Supported	Yes	16-wide leaf intersection

If GPU mode is requested with HKDTree or scalar LBVH, the plugin substitutes the GPU-uploadable LBVH SIMD8 native builder.

Update Type	Use
Static	Level geometry that never moves
Dynamic	Transform-only doors and props; updates the TLAS instance
Refit	Skinned Mesh vertices whose topology stays fixed
Rebuild	Geometry whose triangle topology changes and needs a new BVH

To follow skinned animation, enable both Update Type = Refit and Sync Skinned Vertices On Tick = true. The plugin uploads current-pose vertices and refits the native mesh. This performs CPU skinning and a vertex upload every tick, so limit the number and LOD of updated objects.

Objects with the same Static Mesh, LOD, material mapping, and BVH settings share one native BVH. The Skinned Mesh cache key includes the component path, so distinct poses do not overwrite the same native mesh.

Public methods

Method	Behavior
RegisterNativeObject()	Registers the parent geometry in the native scene
UnregisterNativeObject()	Removes the native object registration
SyncNativeTransform()	Immediately applies the parent transform
RefreshNativeMesh()	Reuploads geometry after a mesh or material change
SyncMaterialsFromParent() (Auto Set Materials)	Automatically matches presets using parent render material names and aliases.
SetMaterialPresetIndex(Slot, Preset)	Changes one slot's preset index
SetMaterialPresetForAllSlots(Preset)	Applies one preset to every slot
SetUpdateType(Type)	Changes the native object update policy
GetResolvedMeshLodIndex()	Returns the LOD index actually uploaded
GetUploadedTriangleCount()	Returns the triangle count from the last upload
IsRegistered()	Returns whether both native object and mesh are valid
GetNativeObjectId(), GetNativeMeshId()	Returns registered native IDs; -1 when unregistered.
GetTargetMeshComponent()	Returns the immediate parent mesh component used for registration.
static IsGpuCompatibleBvhType(ESoundTracingBvhType InBvhType)	Returns true for SIMD LBVH variants.
GetBvhMaxDepth(), GetPrimitivesPerLeafNode(), GetBvhType()	Queries BVH settings in C++.
GetSoundMaterialSlots()	Returns a read-only reference to acoustic material slots in C++.
BuildNativeBvhDebugLineSegments(TArray<FVector>& OutLocalLinePoints)	Retrieves BVH debug segments in local coordinates from C++.
ShouldVisualizeBvh(), GetBvhVisualizationColor()	Editor-only C++ methods for querying BVH visibility and color.

Acoustic materials and Transmission

Detail Properties

The default library is Content/STData/Material/SoundTraceMaterialPresetLibrary.uasset in the plugin. It currently contains the same 22-material table as the Unity and Web SDKs.

Create a custom library from Sounds > SoundTracing > SoundTracing Material Preset Library in the Content Browser. A new asset copies the active default library. Select it under Material Preset Library in Project Settings and restart the Editor.

Field	Description
Presets	Acoustic material list.
Display Name, Aliases	Display name and aliases used to match render materials automatically.
Material Index	Native material index, maintained in library list order.
Scattering	0..1. Balance between specular reflection and scattering.
Reflection, Absorption, Transmission	Eight-band reflection, absorption, and transmission energy coefficients, each in 0..1.
Transmission Model	Selects Surface or Solid Distance.
Thickness to -30 dB (m)	Per-band attenuation reference distance for Solid Distance.
ResetToBundledJson	Restores the preset list from bundled JSON.

Band centers are 67.5, 125, 250, 500, 1000, 2000, 4000, and 8000 Hz. Click or drag the Editor band graph to edit values.

Transmission Model

Model	Input	Geometry requirement
Surface	Per-band energy coefficient 0..1 remaining after one surface crossing	Works with open faces and thin surfaces
Solid Distance	Per-band reference distance in meters at which transmitted energy reaches -30 dB	Requires a closed volume with consistent face orientation

A Solid Distance value is not the object's actual thickness. Runtime measures the distance travelled inside geometry and attenuates it against this reference distance. 0 completely blocks the band.

ImportFromJson, ExportToJson, and ResetToBundledJson use the shared Unity/Web soundMaterial.json format. A missing transmissionDistanceToMinus30DbMeters means Surface; eight valid values mean Solid Distance.

Public methods

Method	Return value / behavior
NormalizePresets()	Normalizes preset indices and per-band values.
GetPresetCount()	Returns the number of presets.
FindBestPresetIndexByName(const FString& RenderMaterialName)	Finds a preset index matching the render material name or aliases.
GetPresetDisplayName(int32 PresetIndex)	Returns the preset's display name.
ResetToBundledJson()	Replaces the list with bundled JSON. Call through the editor button or C++.
ExportToJson()	Returns a string in soundMaterial.json format.
ImportFromJson(const FString& JsonText)	Replaces the list from JSON. Returns false if no materials can be parsed.
FindBestPresetIndex(const UMaterialInterface* RenderMaterial)	Finds a preset matching the render material in C++.
FindPresetIndexByToken(const FString& Token, int32 FallbackIndex)	Finds a preset by token in C++, using the fallback index if no match exists.
static GetFrequencyBandCentersHz()	Returns a read-only reference to the eight center frequencies in C++.
static LoadDefaultLibrary()	Loads the project-selected or default library in C++.
static MakeFallbackPresets()	Creates the fallback preset array in C++.
static ParseSoundMaterialJson(const FString& JsonText, TArray<FSoundTracingMaterialPreset>& OutPresets)	Parses JSON into a preset array in C++.
static SerializeSoundMaterialJson(const TArray<FSoundTracingMaterialPreset>& InPresets)	Serializes a preset array to a JSON string in C++.
static LoadBundledJson(FString& OutJson)	Reads bundled JSON into a string and returns success in C++.

Public properties

Property	Type / access	Description
Presets	TArray<FSoundTracingMaterialPreset> / read/write	Preset array stored in the library.
FrequencyBandCount	static constexpr int32 / constant	Number of frequency bands: 8.

Sound Path Visualizer

Search for Sound Tracing Path Visualizer in the Actor's Add Component menu and add it. This C++ component creates the Niagara component needed for path rendering.

Add SoundTracingPathVisualizerComponent to an Actor to display the latest propagation frame as Niagara line segments.

:::note Bundled NS_Arrow asset paths Niagara System Asset uses the plugin's bundled NS_Arrow asset. Its default path is /SoundTracing/FX/NS_Arrow.NS_Arrow. The associated material is /SoundTracing/Materials/MAT_ArrowLine. Keep both Content/FX and Content/Materials when installing or moving the plugin, and check that the Niagara system and material references remain valid. :::

Detail Properties

Field	Default	Description
Visualization Enabled	On	Enables path display
Refresh Interval Ms	50	Minimum visualization refresh interval; does not affect propagation
Max Visualized Paths	1024	Maximum displayed paths, range 16..5000
Path Alpha Intensity	0.5	Segment alpha strength, range 0.01..2.0
Niagara System	Empty	Uses the plugin's default Niagara system when empty

Colors are Direct=red, Reflection=orange, Diffraction=green, Transmission=cyan, and Reverb=purple. Disable visualization for performance measurements.

Public methods

Method	Return value / behavior
SetVisualizationEnabled(bool bEnabled)	Enables or disables path display at runtime.
GetActivePathCount()	int32. Returns the current number of paths retained by the visualizer.

Sample demos

The three Unity SDK samples are also available as Unreal levels and Blueprints. In the Content Browser, enable Settings > Show Plugin Content, open a level under SoundTracing Content > Samples > Maps, and press Play. Each level has its own GameMode and camera, and audio starts automatically.

All three demos share these controls:

Button	Keyboard	Action
Play	P	Plays the music or conversation from the beginning.
Pause / Resume	Space	Pauses or resumes playback.
Stop	X	Stops playback.
Reset	R	Restarts the demo. Sample 02 also resets positions and the acoustic material.
Show / hide UI	—	Hides or shows the control panel on the left.

Levels are in /SoundTracing/Samples/Maps, behavior and UI Blueprints are in /SoundTracing/Samples/Blueprints, and per-source settings are in /SoundTracing/Samples/Audio/Settings.

ST_SampleScene01

Explore basic spatial audio with one source and a fixed listener in a 10 m room. Listen to the music and examine how room geometry, acoustic materials, and visualized paths relate to the sound.

Use SoundTrace / Unreal dry to switch between SoundTrace output and Unreal's regular output. Both outputs stay synchronized, so you can compare spatial cues and reverberation without restarting the music.

The behavior Blueprint is BP_Sample01, and the UI is WBP_Sample01.

ST_SampleScene02

Move eight sources arranged as left/right guitar, bass, drum, and synth pairs, along with the headphone-shaped listener, in a top-down view. Quartz synchronizes playback so you can compare directionality and room response while changing positions and acoustic materials.

Control	Action
Drag an icon	Use the left mouse button or the first touch to move a source or listener. Its height and the initial grab offset are preserved.
Mirror: ON / OFF	Moves the opposite source symmetrically while dragging. Enabling it does not immediately realign positions.
SoundTrace / Unreal dry	Switches between SoundTrace output and Unreal's regular output while preserving playback position.
Acoustic material (floor + dome)	Changes the acoustic material of both the floor and dome. The initial preset is Glass; this setting is separate from their visual materials.
Reset / R	Restores source/listener positions, Mirror: ON, and the Glass acoustic material, then restarts playback.
Record / save / F9	Starts recording the output. Press again while recording to save a WAV and end the recording.
Save WAV / F10	Saves the active recording as a WAV and ends it.

Recordings are saved with unique filenames in the project's Saved/BouncedWavFiles folder. Saving a WAV in Unreal ends the current recording, unlike Unity's snapshot export during recording. Start dragging inside an icon: drags starting over the UI or within 24 px of the screen edge are ignored.

The behavior Blueprint is BP_Sample02, and the UI is WBP_Sample02.

ST_SampleScene03

Two sources alternate through eight male and female voice clips in a loop, in a space with walls and partitions. Move the listener with the camera to hear changes in occlusion, reflections, HRTF directionality, and room reverberation.

Control	Action
WASD	Moves forward, backward, left, and right.
Q / E	Moves down/up.
Mouse movement	Rotates the view.
Esc	Releases mouse look to interact with the UI. Use Shift+F1 to release the cursor in editor PIE.
Right mouse click	Resumes mouse look.
Reset / R	Restarts the conversation from the beginning at the current camera position.

In editor PIE, Esc ends the play session. The behavior Blueprint is BP_Sample03, and the UI is WBP_Sample03.

Troubleshooting tips

Symptom	Check
Plugin missing from Audio selection	Check SoundTracing activation, the C++ module build, target-platform Audio settings, and editor restart.
Native library load or ABI error	Use the plugin and native library from the same SDK distribution, and ensure ThirdParty files are packaged.
No sound or no spatialization	Check the mono Sound Wave, Audio Component playback, both global Audio plugins, Spatialization activation, and asset assignments.
Source settings not applied	Check the Audio Component → Sound Attenuation → SoundTracing Source Settings chain and Override Attenuation.
Unexpected listener position	With Drive Listener Transform enabled, the component drives the position; otherwise, the Unreal audio listener is used.
Listener replacement warning	Keep one active Listener Component per World.
Geometry not reflected in audio	Check that the Object Component is an immediate child of a supported mesh and that mesh data and triangles exist.
Skinned animation not reflected	Check Update Type = Refit, Sync Skinned Vertices On Tick = true, and LOD.
GPU not active	Check GetGpuBackendStatus() and Output Log. On Windows, deploy exaSound.dll and webgpu_dawn.dll together.
Paths not visible	Check Niagara activation, Visualization Enabled, source/listener path settings, and GetLastValidPathCount().
Pitch spike after teleporting	Call the Listener or Subsystem's ResetMotionState() immediately after changing position.
Dropouts with many sources	Adjust propagation and audio buffers in the order below.

Dropouts with many sources

1. Set Propagation Interval (ms) to 50 to reduce propagation request frequency.
2. Lower Source Ray Resolution Cap to 8..16, or set the source's Ray Preset to Fast.
3. Start with listener settings Quality Preset = Fast, HRTF Path Budget = 1, and Render Band Tier = Merged4.
4. Set Propagation Thread Count to 2..3 and restart to leave execution time for game, render, and audio threads.
5. Set Buffers To Enqueue to at least 2. Adjust Callback Buffer Frame Size if needed.
6. Disable path visualization and change one setting at a time when comparing results.

Runtime status

SoundTracing.Status
SoundTracing.DumpGpuPropagationStats

SoundTracing.Status reports the native version, readiness, processing backend, object/path counts, and listener settings. SoundTracing.DumpGpuPropagationStats reports GPU execution and CPU fallback statistics.

Next

• SDK overview
• Unity SDK
• Web SDK
• STCoreV2