Patrick F. Spear » Posts

Over on Slashdot right now, there’s an article that’s been posted about software developed by Dartmouth researchers that uses your phone’s mic to learn what’s happening around you at any given moment. The program runs on the iPhone, and the idea behind it is pretty simple- it allows the phone to determine what you’re doing by listening in on the specific characteristics of the ambient noise. So it can tell if you’re in the car, biking, at home or in the office just by listening to what’s going on around you. Given that information, the phone can do all sorts of handy things, like increase the ringer volume, send all calls to voicemail, and so on.

The basic idea here is that the phone- carried by a person- will be able to sense its location and behave appropriately. But presumably, the same technology could if the factors are inverted- why not make the location fixed, so the the program is instead determining who is present and what is happening there, so that the space behave appropriately as well?

A kitchen/den could determine who is in a room (from voice tonalities), if a meal is being prepared (from the sound of water running, chopping, or packages being opened), or there is a social gathering. Based on this this information, the selection of music could be changed or light levels set appropriately- down the road, even the arrangement of furniture and partitions could be changed.

The idea could easily apply to offices, halls, meetings rooms, restaurants; the possibilities are actually pretty staggering. Sound is incredibly characteristic of an environment, and you can extract a lot of information from any space’s ambient noises. I think this is a great example of how good interactive architecture doesn’t have to have all sorts of fancy data inputs- instead, it’s better to have fancy AI analyzing more standard data.