Signal Noise Analyzer 📡

A small, ECE-focused project that generates a clean sine signal, adds noise, applies a simple low-pass filter (moving average), and computes Signal-to-Noise Ratio (SNR) before and after filtering. It also visualizes the clean, noisy, and filtered signals.

Features

• Generate sine waves with configurable frequency, duration, and sample rate.
• Add AWGN (additive white Gaussian noise) for a target SNR (in dB) or a custom noise standard deviation.
• Apply a moving average low-pass filter (NumPy-only, no SciPy).
• Compute SNR (in dB) between the clean reference and any processed signal.
• Visualize:
  • Time-domain signals (clean / noisy / filtered)
  • Magnitude spectrum via FFT for noisy and filtered signals

Why this project?

• It’s branch-specific (ECE) and great for interviews/portfolios.
• Shows understanding of DSP basics (noise, filtering, SNR).
• Uses only NumPy + Matplotlib, so it’s easy to run anywhere.

Quick Start

# (Optional) create venv and install
pip install -r requirements.txt

# Run with defaults
python signal_noise_analyzer.py

# Customize parameters
python signal_noise_analyzer.py --f0 5 --fs 500 --duration 2.0 --snr_db 0 --window 15 --save_plots

Output images (PNG) are saved to outputs/ when --save_plots is used.

SNR Definition

We compute SNR (dB) between a reference (clean) signal x and a test signal y as: [ \text{SNR} = 10 \log_{10} \frac{\sum x^2}{\sum (y - x)^2 + \epsilon} ] \epsilon avoids division by zero.

Roadmap / Extensions

• Replace moving average with FIR windowed-sinc or IIR (Butterworth) filter.
• Add Savitzky–Golay smoothing for better shape preservation.
• Add GUI sliders (streamlit) for interactive exploration.
• Support multiple wave types (square/triangle) and multi-tone signals.

---

Dataset Preview

Image	Description