Golay C(23,12) — Python implementation

This repository contains a practical implementation of the binary Golay C(23,12) code and an end-to-end pipeline that: reads 24-bit BMP images, splits the byte stream into 12-bit payloads, encodes them to 23-bit codewords, simulates transmission through a Binary Symmetric Channel (BSC), decodes with IMLD, and reconstructs the image. An experiment harness is included to evaluate Bit Error Rate (BER) and timings.

Requirements

• Python 3.10+ (3.13 recommended)
• Install dependencies:

pip install pillow

Optional (to make a single exe):

pip install pyinstaller
pyinstaller --onefile main.py

Quick start

Run the interactive demo:

python .\main.py

Follow the menu to test:

• 1 — enter a 12-bit vector (interactive)
• 2 — encode/decode a short text string
• 3 — run the full image pipeline on a 24-bit BMP file

Experiments (batch measurements)

Use experiments.py to run a set of trials over several channel error probabilities p and collect averaged BER and timing results. The script will create a 1 MB dummy file (dummy_1MB.bin) if one isn't present.

Run experiments:

python .\experiments.py

Configurable parameters inside experiments.py:

• p_values — list of channel error probabilities to test
• max_workers — number of worker processes used for parallel stages
• num_runs — number of independent runs per p (default in the script: 10)

Outputs produced by experiments:

• experiments_results.csv — row per run with timings and error counts
• experiments_results_avg.csv — per-p averages across num_runs

Notes & troubleshooting (Windows)

• The code uses multiprocessing and spawns worker processes via ProcessPoolExecutor. On Windows the spawn overhead and antivirus (e.g., Windows Defender) can increase timing variance. If you see large fluctuations, try lowering max_workers or reusing a long-lived pool.
• The script calls multiprocessing.freeze_support() to be compatible with frozen executables and Windows spawn semantics.

Files of interest

• main.py — interactive demo and image pipeline
• functions.py — core encode/decode/channel/packing helpers
• experiments.py — experiment harness that saves CSVs
• report.tex — LaTeX report (includes PGFPlots templates using experiments_results_avg.csv)