AudioForge — Documentation

Overview

AudioForge loads audio at runtime — from disk, from a byte buffer, from a downloaded file or from raw PCM — and hands you a USoundWave you can play immediately. No cooking step, no reimport, no editor round-trip.

It also records from input devices, streams audio in as it arrives, exports back out to file or buffer, transcodes between formats, and detects speech with a voice-activity detector.

Import MP3, WAV, FLAC, OGG Vorbis, OGG Opus, BINK and RAW PCM
Stream audio in progressively while it plays
Capture from any available audio input device
Export to file or buffer in any supported format
Transcode between formats without going through a sound wave
Analyse speech boundaries with the built-in VAD
Everything is exposed to Blueprints and C++, and every long operation is async

Requirements

Requirement	Details
	Unreal Engine	4.27 and 5.8 (verified clean on both)
	Modules	`AudioForge` (Runtime), `AudioForgeEditor` (Editor)
	Dependencies	None — self-contained
	Language	Blueprints and/or C++
	Platforms	Win64 — built and verified. The code carries no Win64-only restriction, but other targets have not been verified by us.

Installation

Copy the AudioForge folder into your project's Plugins/ directory (create it if it does not exist).
C++ projects: right-click the .uproject → Generate Visual Studio project files, then build. Blueprint-only projects: just launch the editor — you will be prompted to build the plugin once.
Open Edit → Plugins, search for AudioForge, and confirm it is enabled. Restart the editor if prompted.
To use it from C++, add "AudioForge" to PublicDependencyModuleNames in your module's .Build.cs.

Verifying the install Drop an Import Audio From File node into any Blueprint. If it resolves, the plugin is loaded and ready.

Quick start

Import a file and play it (Blueprint)

Call Create AudioForge to get an importer object.
Bind On Progress and On Result.
Call Import Audio From File with a path and a format (leave it on Determine format automatically if unsure).
In On Result, take the returned sound wave and feed it to a normal Play Sound 2D / audio component.

The same thing in C++

UAudioForgeLibrary* Importer = UAudioForgeLibrary::CreateAudioForge();

Importer->OnResultNative.AddWeakLambda(this,
    [](UAudioForgeLibrary*, UImportedSoundWave* Wave, ERuntimeImportStatus Status)
{
    if (Status == ERuntimeImportStatus::SuccessfulImport)
    {
        UGameplayStatics::PlaySound2D(GWorld, Wave);
    }
});

Importer->ImportAudioFromFile(TEXT("C:/Audio/track.mp3"), ERuntimeAudioFormat::Auto);

Keep a reference The importer and the resulting sound wave are UObjects. Store them in a UPROPERTY() (or a Blueprint variable) or the garbage collector may reclaim them mid-playback.

Supported formats

Import	Export	Notes
MP3	Yes	Yes	LAME tag aware, so seeking and duration are accurate
WAV	Yes	Yes	PCM plus the common integer and float encodings
FLAC	Yes	Yes	Lossless
OGG Vorbis	Yes	Yes	Uses the engine's Vorbis codec
OGG Opus	Yes	Yes	Low-latency speech and music
BINK	Yes	Yes	UE 5.0 and newer only
RAW PCM	Yes	Yes	int8, uint8, int16, uint16, int32, uint32, float32

Importing

Type	Description
Create AudioForge	Callable	Create an importer instance.
Import Audio From File	Callable	Import from a path on disk.
Import Audio From Buffer	Callable	Import from an in-memory byte array (downloads, pak data).
Import Audio From RAW File / Buffer	Callable	Import raw PCM with an explicit layout.
Import Audio From Pre Imported Sound	Callable	Import from a `PreImportedSoundAsset`.
On Progress	Event	Percentage callback during a long import.
On Result	Event	Fires with the sound wave and an import status.
Get Audio Header Info From File / Buffer	Callable	Read duration, channels and sample rate without decoding.
Scan Directory For Audio Files	Callable	Enumerate importable audio in a folder.
Get Audio Format / Get Audio Formats	Pure	Detect a format, or list everything supported.

Playback and sound waves

Type	Description
UImportedSoundWave	Class	The sound wave produced by an import; plays like any other.
UStreamingSoundWave	Class	Append audio while it is already playing.
UCapturableSoundWave	Class	Records from an audio input device.
USynthBasedSoundWave	Class	Procedurally generated audio.
Set Looping / Set Volume / Set Pitch	Callable	Standard playback controls.
Rewind Playback Time / Stop Playback	Callable	Seek or stop.
Get Playback Time / Get Playback Percentage	Pure	Current position, in seconds or percent.
Get Duration Const / Get Sample Rate / Get Num Of Channels	Pure	Wave properties.
Is Playing / Is Playback Finished	Pure	Playback state.
On Audio Playback Finished	Event	Fires when the wave reaches the end.
Set Subtitles	Callable	Attach subtitle cues to the wave.
Release Memory	Callable	Free the decoded PCM data early.

Recording and streaming

Type	Description
Create Capturable Sound Wave	Callable	Make a wave backed by an input device.
Get Available Audio Input Devices	Callable	Enumerate microphones and line inputs.
Start Capture / Stop Capture	Callable	Begin or end recording.
Is Capturing	Pure	Recording state.
Create Streaming Sound Wave	Callable	Make a wave you can append to.
Append Audio Data From Encoded	Callable	Push encoded bytes as they arrive.
Append Audio Data From RAW	Callable	Push raw PCM as it arrives.
Pre Allocate Audio Data	Callable	Reserve buffer space up front to avoid reallocation.

Exporting and transcoding

Type	Description
Export Sound Wave To File / Buffer	Callable	Write a wave out in any supported format.
Export Sound Wave To RAW File / Buffer	Callable	Write raw PCM.
Transcode Encoded Data From File / Buffer	Callable	Convert format to format directly.
Transcode RAW Data From File / Buffer	Callable	Convert between raw PCM layouts.
Resample Sound Wave	Callable	Change sample rate.
Mix Sound Wave Channels	Callable	Change channel count.
Reverse Audio Buffer	Callable	Reverse PCM data in place.
Duplicate Sound Wave	Callable	Deep-copy a wave.
Get PCM Buffer Copy	Callable	Take a copy of the decoded samples.

Voice activity detection

The VAD reports when speech starts and stops in a stream, so you can gate recording, drive push-to-talk, or trigger transcription only when someone is actually talking.

Type	Description
Toggle VAD	Callable	Enable or disable detection on a wave.
Set VAD Mode	Callable	Aggressiveness: `Quality` → `VeryAggressive`.
Set Minimum Speech Duration	Callable	Ignore bursts shorter than this.
Set Silence Duration	Callable	Silence needed before speech is considered over.
Process VAD / Reset VAD	Callable	Run detection manually, or clear its state.
On Speech Started / On Speech Ended	Event	Fire at the boundaries of detected speech.

Configuration

Build-time switches live at the top of Source/AudioForge/AudioForge.Build.cs. Flip one and rebuild.

Default	Effect
bEnableCaptureInputSupport	`true`	Recording from input devices. Windows and Mac.
bEnableVADSupport	`true`	Voice activity detection.
bEnableFileOperationSupport	`true`	Import/export that touches the filesystem.
bEnableMetaSoundSupport	`false`	MetaSounds nodes. Requires UE 5.3 or newer.
bEnableBinkSupport	UE 5+	BINK audio. Automatically off on 4.27.
bUseDrMp3	`false`	Use dr_mp3 instead of minimp3 for MP3 decoding.

Runtime settings Project-level defaults live in Config/DefaultAudioForge.ini and appear under Project Settings.

Troubleshooting

Symptom	Cause and fix
	Import returns a failure status	Check the status value in On Result. `FailedToReadAudioDataArray` means the path or buffer was unreadable; `InvalidAudioFormat` means the data did not match the format you specified — try Determine format automatically.
	Sound plays once then goes silent	The sound wave was garbage collected. Hold it in a `UPROPERTY()` or Blueprint variable.
	No input devices listed	Capture support is Windows/Mac only, and the OS must grant microphone permission to the editor or packaged app.
	MetaSounds nodes missing	`bEnableMetaSoundSupport` is `false` by default and requires UE 5.3+. Enable it and rebuild.
	Long import hitches the game thread	Import is already async — make sure you are reacting to On Result rather than blocking on the call.

Credits

All licence notices, including third-party components, ship at Source/ThirdParty/LICENSE-ThirdParty.txt.

AudioForgeRuntime audio import, capture and export

Overview

Requirements

Installation

Quick start

Import a file and play it (Blueprint)

The same thing in C++

Supported formats

Importing

Playback and sound waves

Recording and streaming

Exporting and transcoding

Voice activity detection

Configuration

Troubleshooting

Credits

AudioForge
Runtime audio import, capture and export