OwnAudioSharp is a cross-platform C# audio framework that gives developers professional-grade audio capabilities with a single, easy-to-use API — no native audio expertise required.
The library is built on a native C++ engine (PortAudio / MiniAudio) backed by a fully managed C# layer. The result is GC-free, dropout-free, real-time audio on Windows, macOS, Linux, Android, and iOS — all from one NuGet package.
| Category | Capability |
|---|---|
| Playback & Mixing | Multi-track sync, real-time tempo/pitch, per-track volume |
| Recording | Low-latency capture with device selection |
| Effects | 15 real-time DSP effects (reverb, EQ, compressor, limiter, …) |
| AI / ML | Vocal separation (HTDemucs 4-stem), reference mastering, chord detection |
| MIDI | Hardware I/O, SMF file read/write, hardware-accurate clock |
| Network | Sample-accurate multi-device sync over LAN |
| Plugins | VST3 effect plugins with cross-platform editor GUI |
| Calibration | SmartMaster speaker calibration with automatic EQ correction |
- Music players and DAWs — synchronized multitrack playback with full effect processing
- Karaoke and stem apps — AI vocal/instrument separation and real-time mixing
- DJ software — tempo/pitch control, network sync, VST3 effects
- Music education tools — chord recognition, MIDI capture, score generation
- Broadcast and podcast tools — reference mastering, dynamic processing, automatic room calibration
- Live performance apps — network-synchronized multi-room or multi-device setups
- Games and interactive audio — low-latency output, real-time effect chains
Three packages are available depending on your platform and feature needs:
| Package | Platforms | AI/ML | Package size |
|---|---|---|---|
OwnAudioSharp |
Windows, Linux, macOS | ✅ Full | ~20 MB |
OwnAudioSharp.Basic |
All platforms | ❌ None | < 7 MB |
AI model files (best, default, karaoke, htdemucs ONNX models, total ~272 MB) are not bundled in the NuGet package. They are downloaded automatically on first use via
VocalRemoverModelManager.DownloadModelAsync()and stored in the user's local application data folder.
dotnet add package OwnAudioSharp # Desktop, full features
dotnet add package OwnAudioSharp.Mobile # Mobile, full features
dotnet add package OwnAudioSharp.Basic # Lightweight, no AI/MLRequirement: .NET 10.0 or later. PortAudio can be installed system-wide for optimal audio quality; otherwise the bundled MiniAudio backend is used automatically.
Play multiple audio files in perfect sync using a shared central clock. Each track supports individual volume, pitch, and tempo control — ideal for DAW-style applications or multitrack players.
Adjust playback speed and pitch independently, in real time, even across multiple tracks simultaneously. Uses the SoundTouch engine under the hood.
Apply reverb, equalizer, compressor, limiter, chorus, delay, distortion, and more to inputs or outputs. Effects are freely combinable and can be inserted per-track or on the master bus.
Load VST3 audio effect plugins and use their native cross-platform editor GUI. Integrates into the effect chain the same way as built-in effects.
Straightforward API for audio capture from any input device with configurable sample rate, buffer size, and channel count.
Measures your speakers using a microphone and automatically corrects the audio output for optimal sound quality. Includes built-in speaker profiles (HiFi, Headphone, Studio, Club, Concert), a 31-band EQ, multiband compression, and a brick-wall limiter.
Good for: room correction, broadcast preparation, professional mastering chains.
Synchronizes audio playback across multiple devices on the local network with sample-accurate precision (< 5 ms on LAN). Zero-configuration with automatic server discovery.
Good for: multi-room audio, live PA setups, museum installations, collaborative production.
Analyzes a reference track and applies its spectral and dynamic characteristics to your audio. Delivers professional mastering results without expensive external plugins.
Good for: music production, podcast finishing, audio restoration.
Separates audio into vocals and instruments (or full 4-stem: vocals, drums, bass, other) using ONNX neural networks. Features the HTDemucs model for high-quality stem isolation with margin-trimming to eliminate chunk-boundary artifacts.
The Multi-Model Pipeline runs several UVR MDX models in parallel and averages their outputs for superior separation quality with fewer artifacts. Available models: htdemucs (4-stem), default, best, karaoke.
Model download: ONNX model files are not bundled in the NuGet package. Call
VocalRemoverModelManager.DownloadModelAsync()to fetch a model on first use — it is saved to the user's local application data folder and reused automatically afterward. Custom model paths are also supported viaSimpleSeparationOptions.ModelPath.
Good for: karaoke creation, remixing, instrumental extraction, stem mastering.
Recognizes major, minor, diminished, augmented, and extended chords (7th through 13th) from audio in real time or offline. Uses a chromagram-based analysis pipeline.
Good for: music transcription, chord chart generation, music education, DJ software.
See the full API reference: OwnAudio/Midi/README.md
AOT-compatible, reflection-free MIDI library supporting Windows (WinMM), macOS (CoreMIDI), and Linux (ALSA rawmidi). Features real-time MIDI input/output, Standard MIDI File (SMF format 0/1) read/write/edit, and a hardware-accurate 24 PPQN MIDI clock.
Good for: sequencers, virtual instruments, MIDI recorders, hardware sync.
OwnAudioSharp uses a two-layer architecture:
Application
└─ OwnaudioNet (high-level API)
└─ AudioEngineWrapper (lock-free, non-blocking)
└─ NativeAudioEngine (C++ PortAudio / MiniAudio)
└─ Audio hardware
- Ownaudio.Core — platform-agnostic interfaces, managed MP3/WAV/FLAC decoders, lock-free ring buffers, SIMD converters, and object pools.
- Ownaudio.Native — cross-platform native engine (PortAudio + MiniAudio fallback) for Windows, Linux, macOS, Android, and iOS.
All blocking engine methods (Initialize, Stop, Send) must be called off the UI thread. The high-level OwnaudioNet API handles threading internally.
Complete API reference, tutorials, and architecture guides are on the official website:
Working example projects are in the OwnAudio/Examples/ directory.
OwnAudioSharp is free and open-source. If you use it in a commercial product or find it saves you time, consider supporting its development:
Your support funds new features, platform improvements, bug fixes, and documentation — and keeps professional audio tooling freely available to the .NET community.
Issues and feature requests: GitHub Issues
See the LICENSE file for details.