[中文：AI 智能人声音高修正工具] OpenTune v1.7.1-1.7.3 STANDALONE VST3 [WiN, MacOSX]

P2P | August 2026  | 537 MB / 1 GB

安装方法：

OpenTune 是一款开源的人声音高修正工具，它将独立编辑器与 VST3 插件工作流程相结合。它并非采用基本的 DSP 移调，而是通过 ONNX Runtime 使用 RMVPE 音高提取和 PC-NSF HiFiGAN 神经声码器来重建修正后的人声，同时力求保留歌手的共振峰。

该编辑器以图形化调音为核心，而非简单的单旋钮自动修正。您可以绘制音高曲线、编辑检测到的音符、使用锚点工具、自动检测调性、修正到常用音阶，并且在宿主程序集成正常的情况下，还可以与支持 ARA2 的宿主程序配合使用。

主要特性
– 神经声码器音高校正，旨在调音和移调过程中保留人声共振峰
– 通过 ONNX Runtime 使用 PC-NSF HiFiGAN 重合成进行 RMVPE 音高提取
– 图形化编辑工作流程，支持手绘、音符编辑、锚点工具和自动调音
– 自动检测音调并校正音阶，支持半音阶、大调、小调、五声音阶、多利亚调式、混合利底亚调式和和声小​​调
– 提供独立应用程序和 VST3 插件版本，兼容宿主软件并支持 ARA2 扩展
– 使用 DirectML（Windows 系统）和 CoreML（macOS 系统）进行 GPU 加速推理
– 支持中文、英文、日文、俄文和西班牙文多语言界面

OpenTune is an open-source vocal pitch correction tool that combines a standalone editor with a VST3 plugin workflow. Instead of basic DSP transposition, it uses RMVPE pitch extraction and a PC-NSF HiFiGAN neural vocoder through ONNX Runtime to regenerate corrected vocals while aiming to preserve the singer’s formants.

The editor is built around graphical tuning rather than one-knob automatic correction. You can draw pitch curves, edit detected notes, use anchor tools, detect key automatically, correct to common scales, and work in ARA2-capable hosts when the host integration behaves correctly.

Key Features
– Neural-vocoder pitch correction designed to preserve vocal formants during tuning and transposition
– RMVPE pitch extraction with PC-NSF HiFiGAN resynthesis through ONNX Runtime
– Graphical editing workflow with hand-drawing, note editing, anchor tools, and automatic tuning
– Automatic key detection plus scale correction for Chromatic, Major, Minor, Pentatonic, Dorian, Mixolydian, and Harmonic Minor modes
– Standalone application and VST3 plugin builds, with ARA2 extension support in compatible hosts
– GPU-accelerated inference using DirectML on Windows and CoreML on macOS
– Multilingual interface support for Chinese, English, Japanese, Russian, and Spanish

🏠 HomePage