fft-viz

A high-quality FFT visualization tool

1. Features

1.1 FFT visualization

• 2D DCT analysis on:
  • Sinusoidal wave image
  • Gray or RGB Gaussian noise image
  • Gray or RGB image from file
• Windowing in pixel-space to mitigate discontinuities on image boundaries
  • 10+ window functions are implemented
• Padding in pixel-space to make sure that the support has zero signals
• 2D power spectral density plot
• Line plots of radial, horizontal, and vertical power spectral density
• Interactive visualization on above features

1.2 FIR filter visualization

• Supported filters:
  • All pass
  • Ideal low-pass
  • Ideal high-pass
  • Butterworth low-pass
  • Butterworth high-pass
  • Chebyshev type I low-pass
  • Chebyshev type I high-pass
  • Chebyshev type II low-pass
  • Chebyshev type II high-pass

2. Requirements

• Python 3.11 or later
• (Optional) CUDA Toolkit 12.0 or later and NVIDIA GPU Used for GPU acceleration. If a compatible GPU is unavailable, the code will fall back to CPU execution with OpenMP

3. Installation

git clone https://github.com/ShineiArakawa/fft_viz.git
cd fft_viz

pip install -r requirements.txt

# Alternatively, if you use 'uv' for dependency management:
# uv sync

4. Usage

Launch the main visualizer with:

cd fft_viz

python fft_viz.py

# Alternatively, if you use 'uv' for dependency management:
# uv run fft_viz.py

# You can also run the FIR (finite impluse response) filter visualizer with:
# python fir_viz.py
# or
# uv run fir_viz.py

The first run may take up to ~20 seconds as it compiles C++/CUDA extensions. Subsequent runs will be much faster thanks to caching.

You can also visualize window functions or FIR filters solely with:

cd fft_viz

python -m lib.windowing

python -m lib.filtering

5. License

This project is licensed under the MIT License. See the LICENSE file for details.