New computer interface: Blow on the screen
Georgia Tech researchers have created a hands-free user interface that uses a microphone to tell where on a computer screen a person is blowing. (From News.com's Underexposed blog.) — Marvel pops comics online, hopes fans pay
Context & Ripple Effects
The Georgia Tech demo lands in the middle of a visible run of alternative-input experiments across tech media: CNET's Underexposed blog had already covered sci-fi-style image navigation in July 2006, and a month before this piece the New Scientist Invention Blog flagged Microsoft's mind-reading work. Labs were clearly probing what could replace the mouse-and-keyboard pairing, using whatever sensor was cheapest to hand.
What distinguishes the Georgia Tech approach is its minimalism: no camera array or electrode headset, just a microphone triangulating breath against screen position. That makes it one of the lowest-cost entries in the hands-free input race, with an obvious audience among users who cannot use their hands.
First-order effects
- Users with limited hand mobility gain a candidate pointing method built entirely from commodity microphone hardware already present on most computers.
- The researchers now face the classic lab-to-product gap for input demos: proving breath control is accurate and fast enough to survive contact with real tasks rather than staged ones.
Second-order effects
- If microphone-based pointing proves viable, it costs nothing extra to ship on existing PCs, putting pressure on hardware-dependent input schemes (cameras, gloves, sensors) that require users to buy peripherals first.
- Accessibility software vendors have a new low-cost interaction channel to evaluate alongside eye-tracking and switch inputs, potentially reshaping which assistive technologies get bundled by default.
Third-order effects
- The broader pattern in this corpus — Microsoft exploring thought detection, MIT students wiring phones into wearable sensing, Georgia Tech using breath — points toward natural-user-interface research becoming a standing category rather than one-off novelty demos.
- If cheap, sensor-light input methods keep maturing, interface design shifts from assuming hands on keyboard to supporting multiple simultaneous body channels, with accessibility needs driving mainstream adoption.
The trend: Mid-2000s input research is fragmenting the mouse-and-keyboard monopoly along body signals — breath, gesture, even neural activity — with low-cost sensors like microphones letting university labs compete directly with corporate R&D.