HachiTune

A lightweight HachiTune built with JUCE framework, featuring FCPE pitch detection and neural vocoder synthesis.

Features • Building • Usage • Shortcuts

Features

• Piano roll interface for visualizing and editing pitch
• FCPE (Fast Context-aware Pitch Estimation) neural pitch detection
• ONNX Runtime vocoder integration with real-time preview
• GPU acceleration support (CUDA, DirectML, CoreML)
• VST3 and AU plugin formats with ARA support
• Multi-language UI (English, 简体中文, 繁體中文, 日本語)
• Undo/redo for pitch modifications
• Real-time audio playback
• Import WAV/MP3/FLAC, export WAV

Prerequisites

• CMake 3.22+
• C++17 compatible compiler
• Git (for submodules)

Platform-specific

Platform	Requirements
Windows	Visual Studio 2019+, Windows SDK
macOS	Xcode with command line tools
Linux	GCC/Clang, ALSA dev libraries

Building

Quick Build

# Clone with submodules
git clone --recursive https://github.com/KCKT0112/HachiTune.git
cd HachiTune

# Build
./build.sh

Manual Build

# Initialize submodules
git submodule update --init --recursive

# Configure
cmake -B build -DCMAKE_BUILD_TYPE=Release

# Build
cmake --build build --config Release

Build Options

Option	Description
ARA_SDK_PATH	Path to ARA SDK for ARA plugin support
ONNXRUNTIME_URL	Custom ONNX Runtime download URL

Output

Build artifacts are in build/PitchEditorPlugin_artefacts/Release/:

• HachiTune.app - Standalone application (macOS)
• HachiTune.vst3 - VST3 plugin
• HachiTune.component - AU plugin (macOS)

Usage

1. Open: Load audio file (WAV/MP3/FLAC)
2. Analyze: Audio is automatically analyzed for pitch
3. Edit:
   • Select mode: Click notes to select, drag to adjust pitch
   • Draw mode: Draw directly on pitch curve
4. Preview: Changes are synthesized in real-time
5. Export: Save modified audio as WAV

Keyboard Shortcuts

Shortcut	Action
Space	Play/Pause
Escape	Stop / Exit draw mode
D	Toggle draw mode
Cmd/Ctrl+Z	Undo
Cmd/Ctrl+Shift+Z	Redo
Cmd/Ctrl+S	Save project
Mouse wheel	Scroll horizontally
Shift+wheel	Scroll vertically
Cmd/Ctrl+wheel	Zoom

Settings

Access via Edit > Settings:

• Language: UI language selection
• Device: Inference device (CPU/CUDA/DirectML/CoreML)
• GPU Device: GPU selection for multi-GPU systems
• Threads: CPU thread count (0 = auto)
• Audio: Output device settings (standalone only)

Project Structure

pitch_editor_juce/
├── Source/
│   ├── Audio/          # Audio engine, pitch detection, vocoder
│   ├── Models/         # Note, Project data models
│   ├── UI/             # UI components
│   ├── Utils/          # Constants, localization, undo manager
│   └── Plugin/         # VST3/AU plugin wrapper
├── Resources/
│   └── lang/           # Localization files
├── third_party/
│   ├── JUCE/           # JUCE framework (submodule)
│   └── ARA_SDK/        # ARA SDK (submodule, optional)
└── CMakeLists.txt