Trying to have a private conversation in public is not an easy task in fact in crowded cities like mine, it’s damn near impossible. But developers at the Karlsruhe Institute of Technology have come up with quite an off beat way of speaking... without actually speaking. Their system uses electromyography to pick up facial contours while speaking and translates them into signals sent to a computer. So even though you’re actually just moving your mouth without actually speaking words out loud the system can recognize the motions and translates them into audio and visual (text) components. Of course it won’t be your own voice but you can’t have everything. If you were to try this out for size right now in it's prototype phase you’d look a little silly walking around with a bunch of wires stuck to your face, but that's also the price of technology in first stages, it never usually looks good on the first try.

Engadget