 |
| ICAT: Nvidia's new weapon in the war against AMD and Intel |
NVIDIA wants you to know that AMD's FidelityFX Super-Resolution technology is just a suggestion. There is nothing like their technology called Deep Learning Super Sampling or DLSS. The GPU can use it to increase a game's frame rate or image quality.
To prove it, Nvidia is trying a new strategy by releasing a very powerful image comparison tool called Nvidia ICAT for free for all to see with their own eyes.
ICAT stands for Image Comparison and Analysis. This tool is a simple program that allows you to view a series of game screenshots or even recorded videos side by side, automatically align the images and move smoothly through each scene, and zoom in and out for each image or videos in the file at the same time displayed.
The tool has a drag and drop interface. With this tool, NVIDIA hopes to show how AMD's ultra-low resolution FidelityFX compares to deep learning ultra-sampling and even NVIDIA's discrete spatial amplification technology.
Spatial optimization technologies such as AMD FidelityFX Super Resolution use static algorithms to play each video frame in the game frame by frame. No special GPU is required for operation.
On the other hand, NVIDIA's deep learning downsampling is a temporary upgrade technology for comparing multiple images.
It takes into account how things move in the video game scene and uses a neural network running entirely on tensor cores to handle all of this that you can find in Nvidia RTX GPUs.
NVIDIA's New Image Comparison Tool
The company also has its own spatial scale called Nvidia Image Scaling. It appears to be hidden in the Nvidia Control Panel. However, the company uses SDKs that support every GPU brand to open on GitHub.
If needed, developers can natively integrate Nvidia Image Scaling into their games, such as AMD's FidelityFX Super Resolution.
The company introduced Nvidia Image Scaling in the GeForce Experience app so you can use it in any game. You can use the NVIDIA overlay to adjust the sharpness slider within the game so you can see the difference.
According to the company, super sampling deep learning, especially the new DLSS 2.3, is better than Nvidia Image Scaling. Sometimes it performs better than the original image.
The company promises that with enough study, the game can look better and run faster than the original DLSS-compatible game.