Researchers at the Massachusetts Institute of Technology have developed a way to extract audio from video silent. It provides a visual image of the sound, and listens to what happened at that point through an algorithm that analyzes the image in the search objects minute vibrations caused by sound waves.
The algorithm picks like virtually indistinguishable vibration and assigns them sound value in accordance with the frequency of vibration. Then he can take the assigned values and reconstruct the sound on their basis.
Sound recovered from standard video recorded at 60 frames per second, inferior to the quality of the recording of high-speed cameras - but it can be used to set a floor speaker in the room, the number of speakers and with sufficient quality even recognize his voice.
The new technology allows you to take the documentary filming and restore it with the voice of a historical figure; get the important details with sound recording surveillance cameras; reconstruct voiceovers old films, optical sound track which was damaged during storage. Of course, all this will require a lot of work but this is the first step in the right direction.
No comments:
Post a Comment