🔊 Upsamplify

Upscale compressed audio like Sony DSEE or Samsung UHQ using AI

🚀 Features

• AI-Powered Upscaling - Restores lost high frequencies (MP3 → Lossless-like)
• Cross-Platform - Works on Windows/macOS/Linux
• Two Interfaces
  • CLI for power users.
  • Streamlit GUI because it's prettier.

⚙️ Installation

Prerequisites

• Python 3.8+
• FFmpeg (for MP3 support)

Install FFmpeg:

Linux (Debian/Ubuntu)

sudo apt install ffmpeg

macOS (Homebrew)

brew install ffmpeg

Windows (Chocolatey)

choco install ffmpeg

1. Clone Repository

git clone https://github.com/yourusername/audio_upscaler.git
cd audio_upscaler

2. Install Dependencies

pip install -r requirements.txt

3. Download AI Models (Auto-download on first run)

python -c "from demucs.pretrained import get_model; get_model(name='htdemucs')"

🖥️ Usage

CLI Version

# Single file
python main.py input.mp3 output.wav

Web GUI

streamlit run app.py

Opens http://localhost:8501 in your browser.

🛠️ Troubleshooting

1. Streamlit Not Found

Fix: Add Python scripts folder to PATH

Linux/macOS

echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.bashrc && source ~/.bashrc

Windows: Add this to PATH manually:

 C:Users\<YOU>\AppData\Roaming\Python\Python3XX\Scripts

2. CUDA Out of Memory

Edit enhance.py:

device = torch.device("cpu")  # Change from "cuda"

3. "Invalid Shape" Errors

• Convert stereo files to mono first using Audacity/FFmpeg
• Trim files to <10 minutes if processing fails

📂 Project Structure

audio_upscaler/
├── input/                  # Default input folder
│   └── sample.mp3          # Example file
├── output/                 # Enhanced audio saved here
├── src/
│   ├── enhance.py          # Core AI processing
│   ├── main.py             # CLI interface
│   └── app.py              # Streamlit GUI
├── requirements.txt        # Dependencies
└── README.md               # This file

❓ FAQ

Q: How much quality improvement can I expect?

A: Typical results:

• MP3 128kbps → Sounds like ~256kbps
• MP3 320kbps → Near-CD quality

Q: Why is the output file so large?

A: WAV is uncompressed. Use ffmpeg to convert to FLAC for smaller files:

ffmpeg -i output.wav -compression_level 12 output.flac

Q: Can I use this commercially?

A: Yes! MIT License allows any use (credit appreciated but not required).