Categories: Retro

Nvidia Uses Deep Learning To Improve Twilight Princess Graphics On Nvidia Shield

Nvidia, the company that makes the chips that power Nintendo Switch and their console Nvidia Shield, has announced something interesting they’ve done with The Legend of Zelda: Twilight Princess on the Chinese Nvidia Shield.

The company recently pushed out an update which dramatically improved the game’s graphics. Nvidia claims this is all thanks to deep learning. Over 4400 textures have been improved after the introduction of the new update:

Top – Before Update | Bottom – After Update

Here are more comparisons:

Nvidia Shield Remaster

Nvidia Shield Remaster After Deep Learning Update

Twilight Princess HD on Wii U

And another:

Nvidia Shield Remaster

Nvidia Shield Remaster After Deep Learning Update

Twilight Princess HD on Wii U

What do you think? Let us know in the comments.

Iggy

Recent Posts

Rumor: Nintendo Switch 2 Stock For Launch Day 3 Times Larger Than Switch In France

Rumors surrounding the initial launch stock for Nintendo Switch 2 have started circulating among the…

5 hours ago

You Can Move Your Animal Crossing: New Horizons Island Over To Switch 2 Via System Transfer

Animal Crossing: New Horizons players should be well aware with the strange way Nintendo handles…

6 hours ago

The Largest Nintendo Switch 2 Game Card Capacity Is 64GB

CDPR, the publisher of Cyberpunk 2077 Ultimate Edition, has confirmed the largest Game Card capacity…

7 hours ago

My Nintendo Store: Switch 2 Edition Physical Releases Have All Content On The Switch 2 Game Card

Yesterday, Marvelous, a third-party publisher, confirmed that Nintendo Switch 2 Edition physical releases have all…

8 hours ago

You Can Rotate The Camera By 360 Degrees In Mario Kart World

A new gameplay feature has been confirmed by players who tried Mario Kart World at…

8 hours ago

Fan Art: What If Nintendo Switch 2 Had Home Menu Themes?

Many Nintendo fans have been requesting Nintendo to add more theme options to Nintendo Switch…

10 hours ago