People were unable to distinguish computer generated sounds from real sounds

Most of the sounds we hear in movies are recreated by specialists. They use the most

different ways:pour water from buckets and break objects. However, researchers are trying to create artificial intelligence that could generate such effects itself. One of these programs has already been tested: people could not distinguish computer-generated sounds from real ones.

Researchers invited for an experimenta small group of people - 53 people. They were asked to listen to different sounds and choose which ones were real and which ones were created by artificial intelligence. 41 out of 53 people could not do it correctly.

Despite this, the researchers note that theirthe algorithm is still far from perfect. So far, the most difficult thing is to synchronize the sounds generated by the AI ​​with the video sequence. The problems that arise here are clearly visible (and audible) in the video below: