DSP Encryption Suite

github.com/something

Overview:

This project consists of two MATLAB scripts using frequency-domain scrambling methods to functionally encrypt an arbitrary signal. The number of bands, the bandwidth of the individual bands and total bandwidth can be adjusted.

Original audio vs Scrambled audio:

How it works

The system encrypts a signal by using a filter to select a certain frequency range, modulates the band by a cosine wave at a specified "key" frequency, and repeats the process for the whole signal. The decryptor simply does the process in reverse; it filters the scrambled signal based on the key, and remodulates it to the original position.

The FIR filter is used to isolate the bands, and the IIR filter is used to eliminate the unwanted sideband from the modulation.

Data Scrambling

The use of scrambling for analog signals is not a new technology. First used during WW2, analog voice scrambling can be decoded by hand using a spectrogram and patience. However, this is not the only use case for this analog method of scrambling.

Digital data signals can be scrambled and descrambled with the same fidelity as analog signals. Adding this layer of encryption on top of an already existing encryption system can obfuscate the protocol used to transmit a given signal.

To demonstrate this, I encoded a base image over SSTV, scrambled the signal and descrambled it to recover the original. The results can be seen below: