🎧 AI-Enhanced Noise Filtering Tool 🔥

February 6, 2025 · View on GitHub

A real-time noise filtering tool that removes background noise from audio using AI and DSP techniques. The application provides a GUI with waveform and spectrum visualization, supports adjustable DSP parameters, and allows recording and saving filtered audio.

🚀 Features

✅ Real-Time Noise Reduction – Uses a deep learning model for denoising.
✅ Adjustable DSP Parameters – Modify cutoff frequency and AI denoising strength.
✅ Waveform Visualization – See real-time changes in the audio signal.
✅ Equalizer (FFT Spectrum Analysis) – Monitor frequency spectrum changes.
✅ Start/Stop Filtering – Control real-time processing with buttons.
✅ Record and Save Audio – Save denoised audio as .wav files.
✅ Dark-Themed GUI – Sleek and modern UI for easy usability.

🛠️ Installation

1️⃣ Install Dependencies

Make sure you have Python 3.8+ installed. Then, install dependencies using:

pip install -r requirements.txt

2️⃣ Run the Application

python main.py

🎛️ How to Use

  1. Start the program by running:

    python main.py

  2. Adjust DSP Parameters: Cutoff Frequency: Controls how much noise is removed. AI Noise Reduction Strength: Adjust the impact of AI denoising. Press "Start Filtering" to begin real-time noise filtering.

  3. Visualize the Audio using: Waveform Graph (Top) – Shows the audio signal in real time. Equalizer (FFT Spectrum) (Bottom) – Displays frequency content. Press "Start Recording" to begin recording filtered audio. Press "Save Recording" to save it as a .wav file. Press "Stop Filtering" to end processing.

🔧 Customization

Use Your Own AI Model – Replace models/pretrained_model.pth with your trained model. Adjust DSP Algorithms – Modify dsp/noise_filter.py to change noise removal techniques. Enhance Visualization – Modify gui.py to add more UI elements.