From the comments:
> Say, we switch our signal at a rate of 440 Hz. Or any rate, as a matter of fact. One would assume that the signal is on for half of the time, and off the other half of the time (which is what’s known as a square wave). But it doesn’t need to be that way. We can also create a signal that’s on for 1% of the time, and off 99% of the time. As long as we switch at 440 Hz, we still get a 440 Hz note. However, we now have a lot of space between the “on” pulses. Within that space, we can render a second train of pulses at a different frequency. The result will be that both frequencies are audible. This gives a very characteristic sound. This is pulse frequency modulation. Funny side node, that’s actually how your brain cells communicate with each other. Anyway, Tim Follin’s early works on ZX Spectrum are typical examples of that.
I wish I could use my relative youth as an excuse here, but I'm afraid I don't quite understand what "on for 1% of the time" means. Like the envelope of the beep is extremely short? And then if you pulse at regular intervals, it creates a sustained tone? And then interleaving differently-spaced pulse sequences creates the illusion of polyphony? Or is it something fancier? Does the ZX have any notion of sample rate or is the signal not even digital?