🎙️ Voice Cloner (Python)

A Python-based voice cloning application that records a voice sample and generates speech in the same voice using an AI-powered Text-to-Speech model.

This project was built as a learning-focused backend/AI application, demonstrating audio processing and speech synthesis using Python.

---

✨ Features

• 🎤 Voice Recording – Record a reference voice sample using a microphone
• 🧠 Voice Cloning – Generate speech that mimics the recorded voice
• 🌍 Multilingual Support – Supports multiple languages
• ⏱️ Adjustable Speed – Control the playback speed of generated speech

---

🛠️ Tech Stack

• Python
• Coqui TTS
• sounddevice
• scipy
• Audio Processing & ML Concepts

---

📦 Installation

Prerequisites

• Python 3.8 or higher
• pip (Python package manager)
• Microphone (for voice recording)
• GPU (optional – improves performance, CPU also works)

---

Setup

1. Install required dependencies

pip install TTS sounddevice scipy

2. First Run Note

• On the first run, the TTS model (~500MB–1GB) will be downloaded automatically

• Ensure:

  • Stable internet connection
  • At least 1GB free disk space

---

🚀 Usage

Step 1: Record Voice Sample

python record_sample.py

• Recording starts after a short delay
• Speak clearly and naturally
• Output file: my_voice.wav
• You can modify recording duration inside the script

---

Step 2: Generate Cloned Voice

python clone_voice.py

• Enter the text you want to convert to speech
• Uses my_voice.wav as the reference
• Output file: output.wav
• Speech speed can be adjusted using the speed parameter

---

📁 Project Structure

Voice-Cloner/
├── record_sample.py    # Records voice sample
├── clone_voice.py      # Generates cloned speech
└── README.md           # Project documentation

---

⚙️ Configuration

record_sample.py

• SAMPLE_RATE – Audio sample rate (default: 16000 Hz)
• DURATION – Recording time in seconds
• output_file – Output voice sample file name

clone_voice.py

• ref_audio_path – Reference voice file path
• out_path – Generated audio output path
• language – Language code (default: "en")
• speed – Speech speed multiplier

---

🧪 Troubleshooting

Module not found

• Run: pip install TTS sounddevice scipy
• Check Python version: python --version

No audio recorded

• Check microphone connection
• Verify system audio permissions

Slow performance

• First run model download is expected
• GPU significantly improves speed
• Ensure sufficient RAM availability

Low audio quality

• Record in a quiet environment
• Speak clearly
• Increase recording duration if needed

---

🧠 Model Information

• Model: tts_models/multilingual/multi-dataset/your_tts
• Languages: Multilingual
• Disk Size: ~500MB–1GB
• Minimum RAM: ~4GB
• GPU: Recommended (optional)

---

📚 Learning Outcomes

• Practical experience with audio data handling
• Understanding voice cloning & speech synthesis
• Hands-on use of Python for backend and AI workflows
• Exposure to machine learning models

---

📜 License & Credits

This project uses the Coqui TTS open-source engine.

• Coqui TTS: https://github.com/coqui-ai/TTS
• sounddevice docs: https://python-sounddevice.readthedocs.io/

---