Nvidia launched turing architecture based rtx 20 series graphics cards. The new cards have a lot of great features. I am gonna talk about what is rtx DLSS and how it works.
What is DLSS?
DLSS stands for deep learning supersampling, based on Nvidia turing architecture technology. It boosts frame rates in games with graphically intensive workloads. It helps gamers to provide high resolutions and settings and a better experience. It provides a better- looking image without using the main processing capabilities of the graphics card. The gamers can view a game of better quality and can experience a higher frame rate.
From a rendered scene multidimensional features are extracted and intelligently combined from multiple frames to get a high-quality final image. DLSS uses few input samples rather than traditional techniques like TAA (Temporal Anti Aliasing) which is owned by Nvidia to avoid difficulties.
The jagged scenes of an object are created by aliasing, and the process known as anti-aliasing makes it worse. In every form, it works by deciding what should fill the gaps between the pixels. DLSS also works on super-sampling, which involves rendering an image of lower resolution to convert it into an image of higher resolution.
Nvidia RTX Series
Nvidia has 4 rtx series cards based on turing architecture. Those are RTX 2060, RTX 2070, RTX 2080, RTX 2080 Ti. RTX 2060 is an entry-level and RTX 2080 Ti is the high-end rtx video card. You can check the rtx 2080 ti cards to find out the best one and why.
How does it work?
The DLSS team first from the selected game extracts the allied frame and a matching perfect frame is generated either by super-sampling or by rendering. They are then transferred to Nvidia’s supercomputers. The DLSS model is then trained by supercomputers to generate aliased inputs to get a high-quality anti-aliasing image which should match the perfect frame. The model is then trained again to generate additional pixels rather than applying AA. GPU renders the full monitor resolution at a higher frame rate by combining both the methods.
DLSS benefit
Each game has its own uniqueness because of its different context, game engine and training time period. The Nvidia’s supercomputer is never turned off and they continue working on the improvement of the quality, even after the game is launched. The further updates are then provided to us by Nvidia’s software updates.
DLSS boosts the frame rates to higher GPU workloads. If a game is already running at a higher GPU framework then, in this case, DLSS not work. On the other hand, if the game is heavily utilizing the GPU, then DLSS will boost the performance. THUS games which run on lower frame rates or higher resolutions get more benefit from the DLSS.
DLSS over TAA
The usage of TAA and DLSS depends on resolution, quality setting and implementation of the game. TAA has been used for many years and it can fall down in certain ways. TAA can suffer from high motion ghosting as it is generated from multiple frames which can be handled by DLSS in a better way.
Drawback
DLSS didn’t turn out to be the most powerful feature of Nvidia’s RTX turing architecture cards as it could have given as better looking games along with better performance. The supporting games for DLSS are also very limited. It is even possible that Microsft’s DirectML can gain more attention.
Within few years DLSS may result as one of the primarily used features due to ease in its installation and dominance of RTX GPU in gamer systems. But if somehow gamers don’t implement it much then it will end up as something very useful or we can say something which will be good to have but not much of importance
FYI, Nvidia is an American technology in Santa Clara, California. Nvidia RTX is used for ray tracing and is a graphics rendering development program created by Nvidia.
