"When you watch Skype or Google Hangouts, often the connection is stuttery and low-resolution and really unpleasant, but the audio is pretty good," said Steve Seitz, co-author of the research and professor at the University of Washington.
"So if you could use the audio to produce much higher-quality video, that would be terrific."
Until now, video lip-syncing involved hours of filming and editing. But the computer program can create a clip with new audio after analysing one hour of speech rather than 14.
It uses artificial intelligence to analyse existing videos online or elsewhere and create realistic simulations from them, taking into account mouth shapes and sound patterns.
"People are particularly sensitive to any areas of your mouth that don't look realistic. If you don't render teeth right or the chin moves at the wrong time, people can spot it right away and it's going to look fake," said Supasorn Suwajanakorn, lead author on the study.
The tool could be used to manipulate videos to create realistic-looking fake clips. But it can only put audio spoken by a person into their mouth.
"You can't just take anyone's voice and turn it into an Obama video," said Seitz. "We very consciously decided against going down the path of putting other people's words into someone's mouth."
It uses artificial intelligence to analyse existing videos online or elsewhere and create realistic simulations from them, taking into account mouth shapes and sound patterns.
"People are particularly sensitive to any areas of your mouth that don't look realistic. If you don't render teeth right or the chin moves at the wrong time, people can spot it right away and it's going to look fake," said Supasorn Suwajanakorn, lead author on the study.
The tool could be used to manipulate videos to create realistic-looking fake clips. But it can only put audio spoken by a person into their mouth.
"You can't just take anyone's voice and turn it into an Obama video," said Seitz. "We very consciously decided against going down the path of putting other people's words into someone's mouth."
A separate organisation earlier this year created an AI that can emulate almost any human voice. The Lyrebird can create new sentence to sound as if they have been spoken by Obama, Donald Trump or Hillary Clinton.
The researchers behind the video manipulation added that their program could also be reversed to make a verification tool. Videos could be fed into the system to check if they are real or have been tampered with, they said.
