If you’ve been looking at the latest and greatest Nvidia RTX graphics cards, you’ve probably heard of DLSS. We have the answers if you’re not sure what it is.
Nvidia has some serious gaming hardware in the form of various graphics cards. Still, in recent years the company has also been working on several software technologies designed to make the most of that hardware.
The company has been focusing heavily on the capabilities of ray tracing and the experience it can bring to games. This technology improves visuals by utilizing realistic lighting, accurate shadows, and diffused light sources. However, it can strain your gaming PC and reduce your FPS count, which is inconvenient for any PC gamer.
This is where DLSS comes into play. This technology allows you to take advantage of amazing visual fidelity while improving performance and maintaining high frame rates to make the most of your fancy gaming monitor.
What is DLSS, and how does it work?
Deep Learning Super Sampling (aka DLSS) is an artificial intelligence technology developed by Nvidia to help improve the performance of your gaming machine (or gaming laptop) even when higher settings are used.
The idea is to put less strain on your GPU by rendering the original image at a lower resolution, then using artificial intelligence to upscale that image to appear to be running at a higher resolution.
Assume you’re playing a visually stunning triple-A title on high settings on a 4K monitor, but the game is running at 1080p and being upscaled to 4K by the AI using DLSS. This means the game can run faster and smoother, resulting in higher FPS without sacrificing visual quality.
The DLSS artificial intelligence is essentially adding extra pixels to produce a fantastic view that is smooth and won’t leave you crying because the FPS counter has dropped.
DLSS isn’t perfect, but it’s improved over time, and Nvidia now has multiple DLSS settings, including performance, balanced, and quality mode. You can switch between these based on your experience, with quality providing the best visual enhancements and performance helping to boost your FPS if it appears to be dropping.
Ray tracing support
We discussed how ray tracing could improve the overall visual effects in the game, but it can also degrade the game’s performance. DLSS counters this by allowing you to keep those settings in place while performing some serious graphics magic to improve overall performance.
This has been seen in a number of games. It makes a bigger difference in some cases than others. When it first came out, Cyberpunk 2077 was particularly taxing on maximum settings. Even with a powerful PC and an RTX 3080, we got 20-30 FPS with DLSS disabled, but everything else turned 11. Turning on DLSS resulted in a nearly doubled frame rate and a much smoother experience.
It is, however, far from perfect. Depending on your playing, you may notice some degradation in overall visual quality when set to performance mode. There may be some blurriness compared to running at 4K without DLSS. However, for the most part, the technology works well, and the trade-off is well worth experimenting with. Read More; Nvidia RTX 4070: Price, Specs, and Release Date
DLSS 2.0 and DLSS 2.1
Nvidia announced DLSS 2.0 in March 2020, an updated version of DLSS that uses a new deep learning neural network that is said to be up to two times faster than DLSS 1.0 because it more efficiently leverages RTX cards’ AI processors, known as Tensor Cores. Because of the faster network, the company can remove any restrictions on supported GPUs, settings, and resolutions.
DLSS 2.0 is also said to improve image quality while promising up to 2-3 times the framerate (in 4K Performance Mode) over the predecessors up to 70% fps boost. Nvidia claims that an RTX 2060 graphics card can run games at maximum settings at a playable framerate when using DLSS 2.0’s 4K Performance Mode. Again, a game must support DLSS 2.0, and an RTX graphics card is required to reap the benefits.
The original DLSS was reportedly limited to about 2x upscaling (though Nvidia has not confirmed this directly), and many games restricted how it could be used. For example, if you have an RTX 2080 Ti or faster GPU in Battlefield V, you can only enable DLSS at 4K — not 1080p or 1440p. This is due to the fact that the overhead of DLSS 1.0 frequently outweighed any potential benefit at lower resolutions and high framerates.
Nvidia released DLSS 2.1 in September 2020, which included an Ultra Performance Mode for super high-resolution gaming (9x upscaling), support for VR games, and dynamic resolution. According to an Nvidia representative, the latter means that “The input buffer’s dimensions can change from frame to frame, but the output buffer’s size remains constant. If the rendering engine supports dynamic resolution, DLSS can be used to perform the necessary display resolution upscale.” It’s worth noting that people frequently refer to both the original DLSS 2.0 and the 2.1 updates as “DLSS 2.0.”
Which games support DLSS?
When DLSS first came out, only a few games supported it, and you couldn’t run it on anything. Since then, DLSS has been improved and released as DLSS 2.0. The company has also attempted to make it more accessible to developers so that it does not need to be trained on specific games in order to perform well. Still, it can instead be more widely integrated into as many games as possible.
DLSS is also supported by Unreal Engine and will work with Unity, implying that it will work with the most popular game engines and a diverse range of games. Read Also; Nvidia RTX 2080 Super vs 3070: Full Comparison
DLSS implementation is still up to the developers, but we’ve seen it in a few games. Nvidia also makes some bold claims about the performance boosts that DLSS offers for the following titles:
- Hitman 3 – the critically acclaimed assassination game- is getting ray tracing and DLSS. In an update, the developers are adding Deep Learning Super Sampling to the game and claiming a two times performance boost while playing at 4K.
- Outriders – with DLSS, every GeForce RTX GPU can hit 60 frames per second at 1440p with NVIDIA DLSS enabled, and the scaling at 4K is up to 73%, turning every GPU from GeForce RTX 3060 Ti and up into a 4K GPU.
- DEATH STRANDING – DLSS enables 100+ FPS at 1440p for all RTX GPUs and can
provide a smooth 60+ FPS at 4K. - F1 2020 – DLSS enables 100+ FPS at 1080p for all RTX GPUs and smooth 60+ FPS at 4K with maximum settings for RTX 2060 Super and higher GPUs.
- Control – DLSS allows gamers to achieve 1440p 60+ FPS with maximum settings and ray tracing enabled on RTX 2070 Super and higher GPUs.
- MechWarrior 5: Mercenaries – DLSS enables 60+ FPS at 4K resolutions for the RTX 2070 Super and higher GPUs and ensures a smooth 80+ FPS at 1440p for all RTX GPUs.
- Deliver Us the Moon – DLSS allows the RTX 2060 and RTX 2060 Super to be played at maximum settings with ray tracing enabled at up to 1440p.
- Wolfenstein: Youngblood – DLSS allows the RTX GPUs to see up to a 2X performance boost at 4K, enabling a much smoother experience.
- Red Dead Redemption 2 – DLSS allows gamers to get over 60 FPS in Red Dead with max settings. Nvidia claims that with a 3060 Ti, you can get over 60 FPS at 1440p, and if you have a 3070 or better, you can get the same high FPS counts at 4K.
- Back 4 Blood – DLSS can improve gaming performance by as much as 46% with this title, Nvidia claims. This enables virtually every GeForce RTX gamer to max out the graphics and get 4K resolution with 60FPS
- Baldur’s Gate 3 – the latest update to Baldur’s Gate 3 has bought DLSS to the game and boosted performance by up to 88% at 4K. Nvidia says if you have an RTX 3060 Ti, you’ll get 4K at 60FPS with this game.
- Battlefield 2042 – Nvidia says that gamers can expect twice the performance at 4K with DLSS on.
The official list of games that support DLSS is ever-expanding but currently includes:
- Amid Evil
- Alan Wake Remastered
- Anthem
- Aron’s Adventure
- Back 4 Blood
- Baldur’s Gate 3
- Battlefield 2042
- Battlefield V
- Bright Memory
- Call Of Duty: Black Ops Cold War
- Call of Duty: Modern Warfare
- Call of Duty: Warzone
- Control
- Crysis Remastered and Crysis Remastered Trilogy
- Cyberpunk 2077
- Death Stranding
- Deliver Us The Moon
- Deep Rock Galactic
- DOOM Eternal
- Enlisted
- Everspace 2
- F1 2020
- F1 22
- Final Fantasy XV
- Fortnite
- Ghostrunner
- Ghost
- Grand Theft Auto: The Trilogy
- Hitman 3
- Hydroneer
- Icarus
- Into The Radius VR
- Iron Conflict
- Justice
- Lego Builders’ Journey
- LEAP
- Loopmacner
- Marvel’s Avengers
- Mechwarrior V: Mercenaries
- Metro Exodus
- Metro Exodus PC Enhanced Edition
- Minecraft
- Monster Hunter: World
- Mount & Blade II Bannerlord
- Mortal Shell
- My Time at Sandrock
- Naraka: Bladepoint
- Nioh 2: The Complete Edition
- No Man’s Sky
- Outriders
- Propnight
- Rainbow Six Siege
- Raji: An Encient Epic
- Red Dead Redemption 2
- Ready or Not
- Redout: Space Assult
- Rust
- Severed Steel
- Scavengers
- Shadow of the Tomb Raider
- Sniper Elite 5
- Soda Crisis
- System Shock Demo
- The Medium
- The Fabled Woods
- Turbo Sloths
- Vampire: The Masquerade – Swansong
- War Thunder
- Warstride Challenges
- Watch Dogs Legion
- Wolfenstein Youngblood
- Wrench
Nvidia claims that enabling DLSS in Rainbow Six Siege can result in a 50% performance boost when playing at 4K. Even with the graphics turned up, the company claims you can get over 100 frames per second.
In the near future, DLSS will also be added to the following games:
- A Plague Tale: Requiem
- Industria
- LEGO Builder’s Journey
- DYING : 1983
- The Ascent
- The Persistence
It’s not just the big gaming titles with DLSS support, either. Numerous indie titles have DLSS capabilities, including:
- Anatomy Of Fear
- Apocalypse: 2.0 Edition
- Beyond Enemy Lines 2
- Bodies of Water VR
- Chivalry 2
- ChronoTecture: The Eprologue
- Cions of Vega
- Death Realm
- DeepStates [VR]
- Exit From
- Fall Balance Ball
- F.I.S.T.: Forged In Shadow Torch
- Frozenheim
- Helios
- Höll Space 5D6
- Last Hope On Earth
- Loverowind
- Mortal Online 2
- Powerslide Legends
- RAZE 2070
- Recall
- REFICUL 666
- Severed Steel
- Soul Dossier
- Sword and Fairy 7
- Swords of Legends Online
- To Hell With It
- Twin Stones: The Journey of Bukka
- Uncrashed: FPV Drone Simulator
- Unknown Woods
- Wakamarina Valley, New Zealand
- Yag
What about DLSS 3?
Along with the RTX-40 series GPUs, Nvidia revealed DLSS 3, which it calls the “next revolution in neural graphics.” This AI technology update combines DLSS Super Resolution, all-new DLSS Frame Generation, and Nvidia Reflex into a convenient package that runs on RTX-40 series graphics cards.
According to Nvidia, this new version of DLSS provides four times the performance of previous generations while also leveraging the power of Nvidia Reflex to ensure optimal responsiveness while gaming.
The promise is faster frame rates, responsive gameplay, and high-quality graphics. You will, however, require one of the RTX-40 series cards. DLSS 3 also requires game developers’ support, but a number of games have already been revealed to support it, including:
- A Plague Tale: Requiem
- Atomic Heart
- Black Myth: Wukong
- Bright Memory: Infinite
- Chernobylite
- Conqueror’s Blade
- Cyberpunk 2077
- Dakar Desert Rally
- Deliver Us Mars
- Destroy All Humans! 2 – Reprobed
- Dying Light 2 Stay Human
- F1 22
- F.I.S.T.: Forged In Shadow Torch
- Frostbite Engine
- HITMAN 3
- Hogwarts Legacy
- ICARUS
- Jurassic World Evolution 2
- Justice
- Loopmancer
- Marauders
- Marvel’s Spider-Man Remastered
- Microsoft Flight Simulator
- Midnight Ghost Hunt
- Mount & Blade II: Bannerlord
- Naraka: Bladepoint
- NVIDIA Omniverse
- NVIDIA Racer RTX
- PERISH
- Portal with RTX
- Ripout
- S.T.A.L.K.E.R. 2: Heart of Chornobyl
- Scathe
- Sword and Fairy 7
- SYNCED
- The Lord of the Rings: Gollum
- The Witcher 3: Wild Hunt
- THRONE AND LIBERTY
- Tower of Fantasy
- Unity
- Unreal Engine 4 & 5
- Warhammer 40,000: Darktide
DLSS in VR
With more recent updates, Nvidia has enabled DLSS support for virtual reality games, including No Man’s Sky, Into The Radius, and Wrench.
You can read it on Nvidia’s official site if you’re keen to learn more about DLSS.
How Does DLSS Work?
The original DLSS and DLSS 2.0 use Nvidia’s NGX supercomputer for AI network training and the Tensor Cores in RTX cards for AI-based rendering.
To enable DLSS 1.0 support in a game, Nvidia had first to train the DLSS AI neural network, a type of AI network known as a convolutional autoencoder, with NGX. It began by displaying thousands of screen captures from the game, each with 64x supersamples anti-aliasing. Nvidia also displayed neural network images without anti-aliasing. The network then compared the images to figure out how to “approximate the quality” of the 64x supersample anti-aliased image with lower quality source frames. The goal was to improve image quality without sacrificing framerate too much.
The AI network would then repeat this process, tweaking its algorithms along the way to eventually match the 64x quality with the base quality images via inference. Nvidia explained in 2018 that the end result was “anti-aliasing approaching the quality of [64x Super Sampled], while avoiding the issues associated with TAA, such as screen-wide blurring, motion-based blur, ghosting, and artifacting on transparencies.”
DLSS also employs what Nvidia refers to as “temporal feedback techniques” (opens in new tab) to ensure sharp detail in-game images and “improved stability from frame to frame.” Temporal feedback is applying motion vectors, which describe the directions in which objects in an image move across frames, to the native/higher resolution output to estimate the appearance of the next frame in advance.
DLSS 2.0 gets its speed boost from an updated AI network that uses Tensor Cores more efficiently, allowing for higher framerates and the removal of GPU, setting, and resolution limitations. In addition, Team Green claims that DLSS 2.0 renders only 25-50% of the pixels (and only 11% of the pixels for DLSS 2.1 Ultra Performance mode) and employs new temporal feedback techniques for even sharper details and improved stability over the original DLSS.
The DLSS 2.0 network, which is also a convolution autoencoder, must still be trained on Nvidia’s NGX supercomputer. According to Nvidia, “low resolution, aliased images rendered by the game engine” and “low resolution, motion vectors from the same images — also generated by the game engine” are involved.
DLSS 2.0 makes use of motion vectors for temporal feedback, which the convolution autoencoder (or DLSS 2.0 network) accomplishes by analyzing “the low resolution current frame and the high resolution previous frame to determine on a pixel-by-pixel basis how to generate a higher quality current frame,” as Nvidia puts it.
The DLSS 2.0 network is trained by comparing the image output to an “ultra-high-quality” reference image rendered offline in 16K resolution (15360 x 8640). Images with differences are sent to the AI network for learning and improvement. This process is repeated by Nvidia’s supercomputer on tens of thousands or even millions of reference images over time, yielding a trained AI network that can reliably produce images with acceptable quality and resolution.
After the AI network’s training for the new game is finished with DLSS and DLSS 2.0, the NGX supercomputer sends the AI models to the Nvidia RTX graphics card via GeForce Game Ready drivers. Then, your GPU can use the AI power of its Tensor Cores to run DLSS 2.0 in real-time alongside the supported game.
Because DLSS 2.0 is a general approach rather than being trained by a single game, the quality of the DLSS 2.0 algorithm can improve over time without requiring a game to include Nvidia updates. The updates are stored in the drivers and may affect all games that use DLSS 2.0.
Nvidia DLSS – Frequently Asked Questions (FAQs)
What does DLSS actually do?
DLSS forces a game to render at a lower resolution (typically 1440p) and then uses its trained AI algorithm to infer what the game would look like if rendered at a higher resolution (typically 4K). It accomplishes this through anti-aliasing effects (likely Nvidia’s own TAA) and automated sharpening.
Should I enable DLSS?
The answer is simple: DLSS uses deep learning to jump through all the hoops of downscaling and upscaling to provide significantly better performance. If your GPU, say an RTX 3060, is unable to achieve 60 FPS in DLSS-enabled titles, simply enable that option in settings and see how well it improves performance.
Does DLSS reduce quality?
As you progress through the list, this affects the game’s rendering resolution, with improved performance but lower image quality. With 2.0, Performance mode provided the greatest improvement, upscaling games from 1080p to 4K.
What graphics cards support DLSS?
With the exception of the shader-core version implemented in Control, DLSS is only available on the GeForce RTX 20, RTX 30, RTX 40, and Quadro RTX series of video cards, which use dedicated AI accelerators known as Tensor Cores.
What is DLSS in gaming?
Deep Learning Super Sampling is an abbreviation for Deep Learning Super Sampling. The technology employs deep learning AI to render games at a lower native resolution and then upscale to your display’s resolution. This gives you the frame rate levels of a lower resolution while maintaining the visual fidelity of a higher display resolution.
