PyVoice

PyVoice is a Python-based speech processing tool that supports both speech-to-text (ASR) and text-to-speech (TTS) using ONNX Runtime. This project refactors and extends an open-source speech project into a production-style CLI application, with cross-platform ONNX Runtime integration.

---

Features

• 🎙 Speech-to-Text (ASR) using Paraformer
  • Supports Chinese and English speech recognition
• 🔊 Text-to-Speech (TTS)
  • Mandarin Chinese and English TTS using MeloTTS
• 🖥 CLI-first design with simple subcommands
• ⚙️ Cross-platform ONNX Runtime integration
• 🧩 Clean and extensible project structure

---

Prerequisites

• Python 3.8+
• pip

Installation

1. Ensure Python 3.8+ is installed

python3 --version  # Should show Python 3.8 or higher

2. Install dependencies

# Install using pip
pip install -r requirements.txt

# Or use pip3
pip3 install -r requirements.txt

# If using virtual environment (recommended)
python3 -m venv venv
source venv/bin/activate  # macOS/Linux
# or
# venv\Scripts\activate  # Windows
pip install -r requirements.txt

3. Verify model files

Ensure the following directories and files exist:

paraformer_weights/
  ├── model.int8.onnx
  ├── tokens.txt
  └── am.mvn

melo_weights/
  ├── model.onnx
  ├── tokens.txt
  └── lexicon.txt

If model files do not exist, download them from Hugging Face (requires git-lfs):

# Install git-lfs if not already installed
brew install git-lfs  # macOS
# or
sudo apt-get install git-lfs  # Linux

git lfs install

# Clone model repository
git clone https://huggingface.co/getcharzp/go-speech ./temp_models

# Move files to correct locations
mv ./temp_models/melo_weights ./melo_weights
mv ./temp_models/paraformer_weights ./paraformer_weights

# Clean up temporary files
rm -rf ./temp_models

---

How to Run

Basic Usage

Run from the project root directory:

# Text-to-Speech (TTS)
python3 main.py tts "<text>" [--out output.wav]

# Speech-to-Text (ASR)
python3 main.py asr <wav-file>

Usage Examples

1. Text-to-Speech

# Chinese TTS (default output to assets/output.wav)
python3 main.py tts "Hello, world!"

# English TTS with specified output file
python3 main.py tts "Hello, world!" --out hello.wav

# Mixed text example
python3 main.py tts "Hello, this is a test!" --out mixed.wav

2. Speech Recognition

# Recognize WAV file
python3 main.py asr assets/output.wav

# Recognize audio file from other location
python3 main.py asr /path/to/your/audio.wav

Help Information

# View help
python3 main.py --help

# View subcommand help
python3 main.py tts --help
python3 main.py asr --help

Notes

1. Audio Format Requirements (ASR):

   • Supports WAV format
   • Automatically converts to 16kHz mono 16-bit
   • If audio format doesn't match, automatic conversion will be performed

2. Output Files (TTS):

   • Default output to assets/output.wav
   • Output directory will be created automatically if it doesn't exist
   • Output format: 44.1kHz mono 16-bit WAV

3. Model Paths:

   • Make sure to run commands from the project root directory
   • Model file paths are relative paths, relative to the project root directory

License

MIT License - see LICENSE file for details.

---

Acknowledgments

This project is based on the open-source project getcharzp/go-speech, with significant refactoring and CLI restructuring.

---