🌟 Overview

Griffith Voice is a cutting-edge tool for translating and synthesizing voice in real-time. It enables seamless communication across language barriers by combining advanced AI models for transcription, translation, and voice synthesis.

Key features:

• 🎙️ Real-time transcription with Whisper
• 🌍 Multi-language translation powered by GPT models
• 🗣️ High-quality voice synthesis using GPT-SoVITS
• 🚀 Streamlit-based web interface for easy interaction
• 📊 Detailed logging and progress tracking

Difference from similar projects: Focus on real-time processing and high-quality voice synthesis.

🎥 Demo

Real-time Translation and Dubbing

Explore Griffith Voice's capabilities in real-time translation and dubbing for multiple languages. Each demo showcases:

• Accurate transcription of speech.
• Seamless translation into multiple languages.
• High-quality voice synthesis for natural-sounding output.

Language Support

Input and Output Languages Support:

🇺🇸 English | 🇯🇵 Japanese | 🇰🇷 Korean | 🇨🇳 Chinese

Input and Output Languages Support Comming Soon:

🇪🇭 عربي | 🇫🇷 French | 🇷🇺 Russian | 🇩🇪 German

Translation supports all languages, while synthesis depends on the chosen TTS model.

Installation

Note: Ensure you have Python 3.10+ and FFmpeg installed.

1. Clone the repository

git clone https://github.com/Si7li/Griffith-Voice.git
cd Griffith-Voice

2. Create and activate a virtual environment

python3 -m venv env
source env/bin/activate

3. Install dependencies

pip install -r requirements.txt

4. Start the application

 run_web_interface.sh

APIs

Griffith Voice supports:

• Transcription: Whisper
• Translation: Google GPT models
• Voice Synthesis: GPT-SoVITS

Why Choose Griffith Voice?

• Optimized for Low VRAM GPUs: Works seamlessly on GPUs with as little as 4GB VRAM, ensuring accessibility for a wide range of users.
• Efficient Memory Usage: Designed to minimize GPU memory consumption without compromising performance.
• Real-time Processing: Delivers fast transcription, translation, and synthesis, making it ideal for live applications.
• Scalable Across Devices: Whether you're using a high-end workstation or a modest setup, the tool adapts to your hardware capabilities.

These features make Griffith Voice a versatile and efficient solution for real-time communication and content creation.

Next Update

• Add Docker implementation
• Add Subtitles that can be exported as .srt/.vtt file

📄 License

This project is licensed under the Apache 2.0 License.

📬 Contact Me

• Submit Issues or Pull Requests on GitHub
• Support me at Patreon
• Email: mohamedkhalil.sahli@enicar.ucar.tn

---

If you find this project helpful, please give it a ⭐️!