Runtime Audio Importer (MetaSound)

Migrating to the MetaSound Audio Pipeline

Projects transitioning from the legacy audio importing setup must remove the standard version of the plugin before enabling the MetaSound edition. This clean substitution ensures that Unreal Engine routes all dynamic sound assets and graph nodes through the MetaSound framework without class namespace collisions. Native support for MetaSounds requires Unreal Engine 5.3 or higher, enabling live audio nodes to plug straight into procedural synthesis graphs, dynamic control buses, and spatialized audio chains.

The plugin integrates smoothly with Unreal Engine's core sound architecture. Every imported track surfaces as an asset that behaves identically to a native Unreal Engine sound wave, preserving parity across all deployment platforms. Developers have access to full runtime API coverage through both Blueprint visual scripting and native C++, making it straightforward to trigger dynamic loads, manage active streams, or read low-level audio data from gameplay code.

Runtime Ingestion and File Transcoding Workflows

Dynamic audio ingestion allows games to pull audio from local storage directories or live memory buffers during active play sessions. The system handles an extensive range of source codecs, decoding MP3, WAV, FLAC, OGG (both Vorbis and Opus encodings), BINK, and RAW PCM data on demand. Games structured through user-generated soundtracks, external speech synthesis engines, or downloadable content packs can parse audio files without requiring pre-cooking through the Unreal Editor.

Format recognition is handled automatically through internal data analysis or file extensions, preventing playback crashes caused by mismatched header information. If an external file lacks an extension or uses an unconventional container, the importer examines the binary structure directly to route it to the proper decoder.

The pipeline functions symmetrically by supporting in-game export and transcoding between codecs:

• Runtime export generates WAV, OGG (Vorbis and Opus), BINK, or RAW PCM audio files directly from active gameplay sessions.
• Real-time transcoding converts loaded streams from one format to another dynamically in memory.
• Directory scanning queries target file paths at runtime to index playable sound files without hardcoded manifests.
• Metadata extraction reads track details, sample rates, durations, and channel configurations immediately upon import.

Specialized Sound Wave Types and Audio Buffer Management

To support diverse runtime performance requirements, the plugin separates dynamic audio handling into three specialized sound wave variants. Each type addresses distinct memory and latency constraints encountered during live playback.

Imported Sound Waves host completely ingested audio assets, giving developers standard playback controls, seeking capabilities, and pitch or volume modulation identical to static sound cues. Streaming Sound Waves provide continuous data streaming governed by dynamic audio buffer management. Through optimized buffer pre-allocation, streaming waves handle incoming network streams or large background tracks without hitching the main thread or causing playback starvation.

Capturable Sound Waves manage real-time hardware input from local recording devices. This variant routes active microphone data into the Unreal audio engine while running real-time speech boundary detection. Across all these types, zero-copy audio data sharing prevents redundant memory allocation when multiple sound instances play the same runtime asset simultaneously. Developers can also tap directly into raw PCM data arrays to manipulate sample values in real time, apply sample rate conversions, or remix channel counts to fit the project's output bus.

Capturing Live Input with Acoustic Echo Cancellation and VAD

Real-time vocal capture introduces acoustic challenges, particularly when players monitor game audio through open desktop speakers. To maintain clean communication channels, the plugin implements WebRTC AEC3 Acoustic Echo Cancellation. This system isolates and strips out game audio output echoing back into the microphone stream, preventing feedback loops and distorted voice transmissions during multiplayer sessions.

Voice Activity Detection operates directly alongside the microphone capture layer. By establishing precise speech boundary detection, the plugin identifies exactly when a player begins and finishes speaking. For production environments subject to background chatter, mechanical keyboard clicks, or ambient noise, integrated Silero VAD provides robust voice detection suitable for noisy conditions. Hardware device permissions are handled automatically per platform, reducing setup overhead when deploying to consoles, mobile devices, or desktop environments. The capture framework also extends to Pixel Streaming setups, allowing server-side instances to capture and process player voice feeds over web streams.

MetaHuman Integration and Cross-Platform Technical Readiness

Dynamic speech analysis opens practical creative opportunities when paired with procedural facial animation frameworks. The plugin provides direct compatibility with Runtime MetaHuman Lip Sync pipelines, allowing runtime audio streams, player voice chat, or dynamic NPC dialogue to drive facial morph targets and phoneme curves across MetaHuman rigs. The same real-time audio analysis feeds into third-party character setups including Character Creator 5, Genesis rigs, and Apple ARKit facial blends.

Cross-platform deployment covers every operating system supported by Unreal Engine. Comprehensive error handling surrounds all runtime operations, reporting broken files, unsupported sample formats, or hardware capture disconnects gracefully through standard Blueprint pins and C++ return delegates. With its memory optimization, format flexibility, and direct hook into Unreal Engine MetaSounds, the plugin offers a production-ready foundation for voice communication, rhythm mechanics, and dynamic audio management.

Explore Similar Assets

• Runtime Audio Importer
• Runtime Speech Recognizer (Real-Time, Offline, AI)
• RealTime Import/Export Plugin
• Runtime DataTable
• Runtime Load FBX