AudioImage

AudioImage is a powerful web-based tool for Audio Spectrogram Art and Digital Steganography, powered by a Flask backend and Svelte 5 frontend. It transforms audio into visual art and allows for secure, invisible file embedding.

Features

• Process Visualizer: A real-time, interactive visualization page that lets you watch and hear the transformation process.
  • Live Waveform: See the audio signal rendered as a glowing, real-time oscilloscope.
  • Step-by-Step Reveal: Watch the spectrogram and steganographic image being constructed layer by layer.
• YouTube Audio Encoder: Directly download audio from YouTube videos (with length validation) and embed it into images seamlessly.
• Audio Art Generation: Convert MP3/AAC audio files into high-resolution visual spectrograms using Python's librosa and matplotlib.
  • GPU Acceleration: Automatically uses torchaudio and CUDA if available for lightning-fast processing.
• Steganography Hider: Hide secret audio or image files inside a "host" PNG image effectively using LSB (Least Significant Bit) encoding.
• Universal Decoder: Decode any image created with AudioImage to retrieve the original hidden files.
• Secure & Private: All processing happens locally on your server; files are cleaned up automatically.

Requirements

• Python 3.12+
• Node.js & npm (or Bun)
• CUDA-enabled GPU (optional, for faster spectrogram generation)

Installation

Method 1: Docker (Recommended for GPU)

The easiest way to run AudioImage with full GPU acceleration is using Docker.

1. Prerequisites: Docker Desktop + NVIDIA Container Toolkit (for GPU).
2. Run:

   docker compose up --build
   

3. Open http://localhost:5000.

Method 2: Manual Setup

1. Backend Setup:

   cd server
   python3 -m venv venv
   source venv/bin/activate
   pip install -r requirements.txt
   

2. Frontend Setup:

   # Root directory
   npm install
   npm run build
   

3. Run:

   cd server
   python app.py
   

Architecture

• Frontend: SvelteKit (SPA mode), Svelte 5 Runes, TailwindCSS.
  • Uses Server-Sent Events (SSE) for real-time progress streaming.
  • Canvas API & Web Audio API for the visualizer.
• Backend: Flask, NumPy, PIL, Librosa, PyTorch (optional).
• Integration: Flask serves the static Svelte build for a unified deployment experience.

Usage Notes

• Supported Audio: MP3, AAC, WAV, FLAC.
• Supported Images: PNG (host/stego).
• Visualizer: Visit the /visualizer page or click the link in the header to see the magic happen in real-time.