Everyday objects, such as a bag of chips or a potted plant, can serve as visual microphone and be used to spy on individuals who may believe their conversations are safe from eavesdroppers.
Sounds a little too sci-fi to be real? Researchers have created an algorithm that converts subtle vibrations caught on film into audible noises and even speech. They claim the technique can lift spoken words off a bag of chips.
Researchers at MIT, Microsoft and Adobe collaborated on the development of the visual microphone algorithm. While a standard 60 frames-per-second (FPS) camera struggles to pick up sounds more discernible than muddled tones, a 6,000 FPS camera can capture music and words using only visual contact with a vibrating object.
"When sound hits an object, it causes that object to vibrate," said Abe Davis, an MIT researcher. "The motion of this vibration creates a subtle visual signature that's usually invisible to the naked eye. In our work we show how using only a video of the object and a suitable processing algorithm, we can extract these minute vibrations. [And we can] partially recover the sounds that produce them, letting us turn everyday visible objects into visual microphones."
It isn't impossible, or too hard, for consumers to purchase a camera that outputs a high enough frame rate to capture discernible audio. Even standard consumer models, rated at 60 FPS, can capture subtle vibrations by employing a technique called "rolling shutters," which pinpoints deviations in frames rather than capturing a high amount of images. Then the vibrations are videotaped and the algorithm is applied to analyze them.
While most citizens won't have to worry about about their wall ornaments literally talking to eavesdroppers, Lauren Weinstein, a Google consultant and policy analyst, said it's highly likely the technology has already been in play for some time.
"[Visual microphones] emphasize the need to keep external windows completely covered during sensitive communications," said Weinstein. "We can safely assume that intelligence agencies -- at a minimum -- have been using this technique for some time."
While some privacy advocates aren't likely to be happy about the research effort, Davis said there were likely better sound-reconstruction tools in existence.
"Big brother won't be able to hear anything that anyone ever says all of a sudden," said Davis. "But it is possible that you could use this to discover sound in situations where you couldn't before. It's just adding one more tool for those forensic applications."
