The Gaming Restoration | Far Cry 3 - Stranded Trailer (Remastered in 5K using AI Machine Learning) [Impressive Results] @TheGamingRestoration | Uploaded 5 years ago | Updated 3 hours ago
Remastered in 5K using ESRGAN [AI Machine Learning]

ESRGAN is an AI Neural Network software that is capable of upscaling and reconstructing
low-resolution images, while retaining most of their elements and details intact.

I used the original uncompressed trailer that Blur Studio released back in 2012. I don’t
remember exactly where did I get the footage, but I believe it was from the time when
websites like Gametrailers used to let you download trailers and gameplay videos in
their original quality.
This time I wanted to upscale a video that was already in HD and had a high bitrate, to
see how ESRGAN and the Manga109 training model would work without obstacles
like compression artifacts, interlacing or blocking.

Since the resolution of the source material was twice as big as any standard PS2-era
video, ESRGAN was able to upscale the footage up to 5K. The sharpness and clarity
of the image starts a bit off (because the trailer begins in the perspective of a video
camera), but once you reach the second half of the trailer, the footage almost looks
like it was rendered natively in 5K.


Original resolution: 1280x544
ESRGAN Upscale: 5120x2179 [2.35:1 Aspect Ratio]

ESRGAN: github.com/xinntao/ESRGAN
Manga109Attempt.pth: http://www.mediafire.com/file/w3jujtm752hvdj1/Manga109Attempt.pth.zip/file
KNLMeansCL: github.com/Khanattila/KNLMeansCL