Warum brauchst du den bpm part? Reicht das nicht ein accelerometer an einem esp um das zu erfassen? Oder Mikro oder vergleichbares

Das klopfen dann über was separates mit Motor und Relais

Interferometry, maybe?

Two microphones, side by side. Signals that come from directly in front (or directly behind, unfortunately) will hit both microphones simultaneously. But sounds coming from the side or any other angle will hit each microphone at slightly different times. If you then have a hardware or software solution to filter out any sound waves that don't match, you've got a highly directional microphone that should work well in any frequency range. (Can also use this setup with 3 microphones if it needs to be 3D and filter out noise from above and below as well.)

There are commercially available setups that do this, but in theory it shouldn't be too hard to build your own, either. I mean, in theory, you could just feed the outputs of both mics through an and-gate transistor into a single output, and that very simple hardware would do the filtering for you -- you'll only get the outputs from when both mics pic up the same sound at the same time.

Ok, I've done experiment myself; it was a dumb idea. Sure I can see the bass beat on smarthone hall sensor as sharp periodic spikes... from few cm away of the coil.

Thing is, any measurable signal outside of coupled system is essentially power loss, which industrial made speakers are, obviously, optimized against. The field loops back into the magnetic drive through the shortest achievable path. Then, it's not your typical radiative EM system, it's low frequency near field transmission - field drops exponentially as you move away from the coil.

Most probably, you'd need to measure long periodic signal, do lots of oversampling, etc. to gen anything. Or build a high fidelity resonator to steal the field from the speaker - essentially, same thing, except the resonator will probably start singing, too.

So I'd give up on magnetic approach; why go for counteroptimized signal when you have optimized acoustic field? Just analyze it - look for sustained harmonic vibrations, fat envelopes so different from human voice barking and foorsteps; EDM are full of those nice sounds.