FSK Modulation and Demodulation: A Robust Python Implementation

This repository provides Python scripts for encoding and decoding Frequency-Shift Keying (FSK) signals. FSK is a digital modulation technique that represents digital data by variations in the frequency of a carrier wave. This implementation prioritizes accuracy and robustness, particularly in noisy environments.

Key Features

• fsk-encode.py: Encodes binary data into an FSK audio signal (WAV file). Customize carrier frequency, frequency deviation, baud rate (bits per second), and sample rate.
• fsk-decode.py: Decodes FSK signals from WAV files, outputting the original binary data. Employs a refined Short-Time Fourier Transform (STFT) based algorithm for accurate frequency analysis. Includes detailed logging for debugging.

Enhanced Decoding Algorithm

This version boasts a significantly improved decoding algorithm compared to previous iterations:

• Robust Peak Detection: A refined peak detection method minimizes the effects of noise and interference on decoding accuracy.
• Precise Bit Synchronization: The decoder utilizes the specified baud rate for precise bit boundary identification, eliminating synchronization errors. Crucially, the baud rate used for decoding must match the encoding baud rate.
• Improved Error Handling: The code includes comprehensive error handling to gracefully manage various scenarios, such as invalid input parameters or file errors.
• Detailed Logging: The decoder provides detailed logging at various levels (DEBUG, INFO, WARNING, ERROR, CRITICAL), allowing for in-depth analysis of the decoding process.

Installation

This project requires several Python libraries. Install them using pip:

pip install numpy scipy wave librosa argparse matplotlib

The following libraries are used:

• NumPy: For numerical operations.
• SciPy: For signal processing functions (e.g., filtering).
• Wave: For WAV file I/O.
• Librosa: For audio analysis (STFT).
• Argparse: For command-line argument parsing.
• Matplotlib: (Optional) For visualizing spectrograms.

Usage

Encoding (fsk-encode.py)

Encode binary data:

python fsk-encode.py "101101001010" --frequency 10000 --deviation 500 --baud-rate 1000 --sample-rate 44100 --output encoded.wav --debug 20

• "101101001010": The binary data to encode.
• --frequency 10000: Carrier frequency (Hz).
• --deviation 500: Frequency deviation (Hz).
• --baud-rate 1000: Baud rate (bits per second).
• --sample-rate 44100: Sample rate (Hz). (Default: 44100 Hz)
• --output encoded.wav: Output WAV file name. (Default: output.wav)
• --debug 20: Sets the logging level (10=DEBUG, 20=INFO, 30=WARNING, 40=ERROR, 50=CRITICAL). Default is 20 (INFO).

Decoding (fsk-decode.py)

Decode a WAV file:

python fsk-decode.py encoded.wav --frequency 10000 --deviation 500 --baud-rate 1000 --debug 10

• encoded.wav: Path to the WAV file.
• --frequency 10000: Carrier frequency (Hz) — must match encoding.
• --deviation 500: Frequency deviation (Hz) — must match encoding.
• --baud-rate 1000: Baud rate (bits per second) — must match encoding.
• --debug 10: Sets the logging level. Use --debug 10 for detailed DEBUG messages.

Example Workflow

1. Encode: python fsk-encode.py "11001100" --frequency 12000 --deviation 1000 --baud-rate 500 --output my_signal.wav --debug 20
2. Decode: python fsk-decode.py my_signal.wav --frequency 12000 --deviation 1000 --baud-rate 500 --debug 10

Contributing

Contributions are welcome! Please open an issue or submit a pull request. Ensure your code adheres to the existing style and includes comprehensive tests.