Boffins at MIT, Microsoft and Adobe have developed an algorithm that can bring audio to silent videos by reading vibrations that are not visible to the naked eye on objects.
In one example, the team was able to recover “intelligible speech” from a video thanks to vibrations that a voice made on a packet of crisps, which was positioned 15 feet from the camera and behind soundproof glass. Other tests successfully reconstructed sound using vibrations from aluminum foil, a glass of water, and a potted plant.
“Sometimes we watch these movies, like James Bond, and we think, ‘This is Hollywood theatrics. It’s not possible to do that. This is ridiculous,'” said Alexei Efros, an associate professor at the University of California at Berkeley. “And suddenly, there you have it. This is totally out of some Hollywood thriller.”
➤ Extracting audio from visual information [MIT] | Via Techmeme
Thumbnail image via veleknez / Shutterstock
Get the TNW newsletter
Get the most important tech news in your inbox each week.