A scene that shows the American showman “Buffalo Bill”, shot more than a hundred years ago and interviewed the leader of the American Indians Oglala Lakota, gained a quality as if it was taken yesterday thanks to artificial intelligence. In very old movies, the motion seemed unnaturally fast because film cameras that recorded using the twist of time used to shoot at a lower frame rate than current cameras. The re-edited images, on the other hand, continue to be black and white, but it does not appear to be accelerated as in other silent films.

Even images from the 1800s could be converted to HD videos

A post shared by My Colorful Past (@my_colorful_past) on Sep 18, 2019 at 11:26am PDT Restoring historical photographs on his website, My Colorful Past, digital artist Matt Loughrey brings movies shot in the late 19th and early 20th century to almost today’s standards. Loughrey uses artificial intelligence to recreate missing visual information between the original frames of the movies. In this way, it can ensure that the movement flows smoothly like in modern film and video. The images that William “Buffalo Bill” spoke with Cody’s leader of the native American tribe Oglala Lakota, also known as “Chief Iron Tail”, Siŋté Máza, were shot in 1914. The original images of his speeches were taken at about 19 FPS with the limited technology at that time. It should be noted that the frame rate is 24 FPS for today’s modern movies and 60 FPS for high definition (HD) videos.

Colorized version of the video that Buffalo Bill speaks with Siŋté Máza in sign language:

A post shared by My Colorful Past (@my_colorful_past) on Jun 1, 2020 at 9:49am PDT One of the reasons why the details in the video created with artificial intelligence look as sharp as in modern movies is high FPS. Loughrey designed an algorithm that created new frames between the original frames of the movie to modernize the “Buffalo Bill” clip. However, Loughrey stated that this algorithm is different from motion interpolation, another video processing technique that duplicates and combines existing frames. The algorithm processed about a minute of Buffalo Bill images in 40 hours and produced thousands of new frames. The result is a high image quality that appears to be real-time at about 60 FPS. Loughrey also reached 71 FPS in a video showing Broadway in New York, shot at 16 FPS in 1896. Despite the 19th century architecture, tools and clothes, the way people move emphasizes the small details that make the scene look like it was taken today.

Refreshing images taken on Broadway:

Loughrey describes his work as different versions of time travel. If you want to review other works restored by the artist, you can access the Instagram account here.

An Artist Converted Nearly 130 Years Of Images To HD Videos With Artificial Intelligence - 65