Digital signals, despite their binary nature, can be represented as a sum of sinusoidal waves with different frequencies - a concept described by Fourier analysis.
This simulation shows how a digital signal can be decomposed into its frequency components. The frequency spectrum displays the amplitude of each frequency component present in the signal.
The red overlay on the time domain shows how adding harmonics progressively reconstructs the original digital signal. As you increase the number of harmonics used, the reconstruction becomes closer to the original square wave, demonstrating Gibbs phenomenon and the theoretical infinite bandwidth needed for perfect square waves.