Deep Learning + Digital Forensics/Exploitation

Supposed an there is an audio recording of the keyboard typing by a person.   From the audio, is it possible to know exactly what have been typed?

From normal ear, it is not possible.   But if you are able to take the audio waves and break it down to its simpler component, it may be possible to see the alphabet.   This is because each key entry have its own unique audio signal distribution as compared with another key.   Every keys will can uniquely identified by the audio signal of the click.

Similarly, listening to the audio signal generated by the ATM machine (password entry pad) the same thing may be said as well:   even if the electronics enforced using the SAME tone generator to generate the sound for ALL keys, the other noises added to each keys may be different from one another, and thus unique characterization of the keys are possible.

And this is where Deep Learning training is highly applicable:   the rules are not codifiable, but somehow it exists.