ErTing - AI-Powered Audio/Video Denoising

A powerful, local-first audio/video denoising tool powered by AI. ErTing uses ModelScope's speech enhancement models to remove noise from your audio and video files, delivering clean, professional-quality output.

ErTing (耳听) -- Hear clearly.

Features

Three Interfaces

• CLI - Full-featured command-line interface with unified flags
• GUI - Desktop Tkinter interface for easy point-and-click operation
• Web - Flask-based web interface for browser-based denoising

Core Capabilities

• AI-powered noise suppression using ModelScope models
• Supports wide range of audio/video formats (MP3, WAV, MP4, AVI, MOV, M4A, FLAC, OGG, WMA, AAC)
• Automatic format conversion to 16kHz WAV for optimal processing
• Configurable output paths and model selection
• JSON output for automation workflows

Requirements

• Python 3.10+
• ModelScope package
• pydub for audio format conversion
• 2GB+ RAM recommended for model loading

Installation

Method 1: Install from PyPI

pip install erting

Method 2: Install from Source

git clone https://github.com/cycleuser/ErTing.git
cd ErTing
pip install -e .

Quick Start

CLI Usage

# Basic denoising
erting input.mp3

# Specify output path
erting -o output_clean.wav input.mp3

# JSON output for scripting
erting --json input.mp3

# Verbose mode
erting -v input.mp3

# Quiet mode (minimal output)
erting -q input.mp3

GUI Usage

erting-gui

Web Interface

erting-web

Then open http://localhost:5001 in your browser.

CLI Options

erting [-V] [-v] [-o PATH] [--json] [-q] input [options]

Positional Arguments:
  input                  Input audio/video file path

Options:
  -V, --version         Show version and exit
  -v, --verbose         Enable verbose output
  -o, --output PATH     Output file path
  --json                Output result as JSON
  -q, --quiet           Suppress non-essential output
  --model MODEL         ModelScope model name (default: iic/speech_zipenhancer_ans_multiloss_16k_base)

Python API

from erting.api import denoise_audio, ToolResult

result = denoise_audio(input_path="input.mp3")
print(result.success)    # True / False
print(result.data)      # {'input_path': ..., 'output_path': ...}
print(result.metadata)  # {'version': '0.1.0', 'model': ...}

Agent Integration (OpenAI Function Calling)

ErTing exposes OpenAI-compatible tools for LLM agents:

from erting.tools import TOOLS, dispatch

# Pass TOOLS to the OpenAI chat completion API
response = client.chat.completions.create(
    model="gpt-4o",
    messages=messages,
    tools=TOOLS,
)

# Dispatch the tool call
result = dispatch(
    tool_call.function.name,
    tool_call.function.arguments,
)

Supported Formats

Format	Extension	Notes
WAV	.wav	Direct processing
MP3	.mp3	Direct processing
MP4	.mp4	Audio extracted
AVI	.avi	Audio extracted
MOV	.mov	Audio extracted
M4A	.m4a	Direct processing
FLAC	.flac	Direct processing
OGG	.ogg	Direct processing
WMA	.wma	Direct processing
AAC	.aac	Direct processing

Project Structure

ErTing/
├── pyproject.toml              # Package metadata & build config
├── README.md                   # This file
├── LICENSE                     # GPL-3.0-or-later
├── erting/
│   ├── __init__.py             # Package init
│   ├── __version__.py          # Version
│   ├── __main__.py             # python -m erting entry
│   ├── core.py                 # Audio processing engine
│   ├── cli.py                  # CLI with unified flags
│   ├── gui.py                  # Tkinter GUI
│   ├── web.py                  # Flask web app
│   ├── api.py                  # ToolResult API
│   ├── tools.py                # OpenAI function-calling tools
│   └── templates/
│       └── index.html           # Web UI template
├── tests/
│   ├── __init__.py
│   ├── conftest.py             # Test fixtures
│   ├── test_core.py            # Core module tests
│   ├── test_api.py              # API tests
│   ├── test_tools.py           # Tools schema tests
│   └── test_cli.py             # CLI integration tests
└── scripts/
    ├── publish.sh               # PyPI publish script
    └── publish.bat              # Windows publish script

Testing

# Run all tests
python -m pytest tests/ -v

# Run specific test file
python -m pytest tests/test_api.py -v

# Run with coverage
python -m pytest tests/ --cov=erting --cov-report=term-missing

Publishing to PyPI

# Linux/macOS
./scripts/publish.sh

# Windows
scripts\publish.bat

Or manually:

rm -rf dist/
python -m build
twine upload dist/*

Development

# Install dev dependencies
pip install -e ".[dev]"

# Run tests
python -m pytest tests/ -v

# Format code
ruff format .

# Lint
ruff check .

# Type check
mypy erting/

License

GPL-3.0-or-later. See LICENSE for details.