Voice Cloning TTS Project / 声音克隆 TTS 项目

Author / 作者: 江景哲 JIANG JINGZHE
Contact / 联系方式: contact@jiangjingzhe.com

🔊 Audio Demo / 效果试听

▶️ Click here to listen to the demo / 点击此处播放试听样本

🇬🇧 English Version

Introduction

This is a simple Python project that uses Coqui TTS (XTTS v2) to clone your voice and perform text-to-speech generation. It is optimized for macOS (Apple Silicon) environments.

Prerequisites

1. Environment Setup:
   The project uses a virtual environment.

   Activate the virtual environment:

   source venv/bin/activate

   Install dependencies (if not already installed):

   pip install -r requirements.txt

2. Recording Voice:
   You can use the included script to record directly:

   python record_audio.py

   Or record manually:

   • Record about 60 seconds of your voice.
   • Speak clearly in a quiet environment.
   • Save as WAV format.
   • Rename the file to my_voice_60s.wav.
   • Place the file in the samples folder.

Usage

Option 1: Single Generation (main.py)

If you only want to generate one sentence:

python main.py

Modify the TEXT_TO_SPEAK variable in main.py to change the text.

Option 2: Interactive Fast Generation (Recommended)

If you want to generate multiple sentences continuously without reloading the model every time:

python interactive_tts.py

1. The model loads once at startup.
2. Type a sentence and press Enter; it will generate and play automatically.
3. Type q to exit.

🐳 Docker Usage

1. Build the Image:

   docker build -t my-tts .

2. Run the Container:
   You need to mount the samples (input) and output (results) directories.

   docker run -it --rm \
     -v "$(pwd)/samples:/app/samples" \
     -v "$(pwd)/output:/app/output" \
     my-tts

   Note: Audio playback will not work inside the container. Please check the output folder for generated files.

⚠️ Security & Ethical Use Disclaimer / 安全与道德使用声明

English:
This project is intended for educational and research purposes only.

• Do not use this software to clone voices without the explicit consent of the speaker.
• Do not use this software to generate content that is illegal, harmful, defamatory, or intended to deceive (e.g., deepfakes for fraud).
• The authors assume no responsibility for any misuse of this software. By using this software, you agree to take full responsibility for your actions.

中文:
本项目仅供教育和研究目的使用。

• 请勿在未获得说话者明确同意的情况下克隆其声音。
• 请勿使用本软件生成非法、有害、诽谤或旨在欺骗的内容（例如用于诈骗的深度伪造）。
• 作者不对本软件的任何滥用行为承担责任。使用本软件即表示您同意为您的一切行为承担全部责任。

📄 License / 许可证

This project is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0).
本项目采用 GNU Affero General Public License v3.0 (AGPL-3.0) 许可证。

See the LICENSE file for details.

🇨🇳 中文版本

简介

这是一个简单的 Python 项目，使用 Coqui TTS (XTTS v2) 来克隆你的声音并进行文字转语音。针对 macOS (Apple Silicon) 环境进行了优化。

准备工作

1. 环境配置:
   项目已配置好虚拟环境。

   激活虚拟环境:

   source venv/bin/activate

   安装依赖 (如果尚未安装):

   pip install -r requirements.txt

2. 录制声音:
   你可以使用自带的录音脚本直接录制：

   python record_audio.py

   或者手动录制：

   • 录制一段大约 60 秒的你的声音。
   • 说话清晰，背景安静。
   • 保存为 WAV 格式。
   • 将文件重命名为 my_voice_60s.wav。
   • 将文件放入 samples 文件夹中。

运行

方式一：单次生成 (main.py)

如果你只想生成一句话：

python main.py

修改 main.py 中的 TEXT_TO_SPEAK 变量来改变文字。

方式二：交互式快速生成 (推荐)

如果你想连续生成多句话，不需要每次都等待模型加载：

python interactive_tts.py

1. 程序启动后会加载一次模型。
2. 然后你可以像聊天一样，输入一句话，回车，它就会立刻生成并自动播放。
3. 输入 q 退出。

🐳 Docker 使用方法

1. 构建镜像:

   docker build -t my-tts .

2. 运行容器:
   你需要挂载 samples (输入) 和 output (输出) 目录。

   docker run -it --rm \
     -v "$(pwd)/samples:/app/samples" \
     -v "$(pwd)/output:/app/output" \
     my-tts

   注意: 容器内无法直接播放音频。请在 output 文件夹中查看生成的文件。