One of my favorite books, Daemon (by Daniel Suarez), has a brilliant programmer named Matthew Sobol, who uses directional high-frequency sound to simulate a voice being spoken right in your ear. The premise is that if you have two sine waves of differing frequencies colliding, they produce a sine wave at the difference of the two (in other words, you could transmit audio at 50,000 and 51,000 Hz, and at the point of intersection, it would seem like 1,000 Hz sound came out).

Talking about it at one of my favorite podcasts, Security Now!, host Steve Gibson gave me an idea - given that there's ubiquitous software now to convert the pitch of sound, why not write software to take a sound file or input from a microphone, split it in two, and increase its frequencies well above audible range? It might allow for the same 'beaming' effect, and even could be used to simulate quality speakers where ones can't be mounted.

It's an experiment I've been having a lot of fun playing around with. I'll post notes and software downloads as they become ready, and hope you'll enjoy them.