DLSS 4: Nvidia's Game-Changing AI Technology Sparks Debate

Nvidia's RTX 50-series could be a huge flop if gamers reject DLSS 4

Nvidia's GeForce RTX 5090 and RTX 5080 are real, and well, they're not as spectacular as many gamers hoped at first glance - at least when you aren't using DLSS 4. I'm worried about what that means for the $550 RTX 5070 PC that gamers seemed so jazzed about, and whether the entire RTX 50-series could wind up being a flop if DLSS 4's fantastic new Multi Frame Generation fails to pick up steam fast. (It should!) And I have the numbers to back up my fears. The "Blackwell"-powered RTX 50-series doesn't deliver tangible improvements in value or raw rendering performance from generation to generation, full stop. The GeForce RTX 5090 brute forces its way atop the benchmark charts thanks to Nvidia cramming it with 33 percent more CUDA cores than the RTX 4090, which helps it achieve 27 percent faster speeds at 4K resolution - for an 18 percent increase in price and much higher power consumption. It's definitely brutal and badass, but meh for a new generation. The RTX 5080 is even less impressive, delivering only 15 percent more performance than the RTX 4080 Super it replaces at 4K, and a paltry 11 percent improvement at 1440p. Plus, it still only has 16GB of GDDR7 memory despite costing $1,000. Meh. Balancing this out, of course, are the massive performance gains provided by Nvidia's new DLSS 4, and more specifically, its Multi Frame Generation feature. MFG inserts up to three AI frames between two "traditionally" rendered frames to send frame rates and visual smoothness soaring. The tightly paced frames it spits out can make even janky games like Star Wars Outlaws feel as snappy as the legendary Doom 2016. As I said in our RTX 5090 review, it's a true, literal game-changer, and it will work in 75 games and apps from day one. But the RTX 5080 shows that Nvidia is willing to bet an entire graphics card's existence on DLSS 4's promise. And the RTX 5080 isn't the only one. Shortly after Nvidia revealed the Blackwell generation, we analyzed the RTX 50-series specs, trying to estimate where "traditional" gaming improvements will fall. We came damn close to nailing it, guessing that the RTX 5090 would wind up roughly 33 percent faster (it wound up at 27 percent) and the 5080 would wind up 10 to 15 percent faster (reality: 11.5 to 15 percent). How'd we do that? Simple: We simply compared their CUDA core counts to their predecessors. Blackwell GPUs are built using the same 4N process at TSMC, so there was no easy uplift coming from a transistor node jump, and instead of prioritizing traditional performance gains, Nvidia focused much of its attention on beefing up the RTX 50-series' AI tensor cores to accommodate DLSS 4's delightful DLSS 4 Multi Frame Generation feature, as well as more future-looking technologies like Neural Rendering and Mega Geometry. Again: DLSS 4 is so much more than just "fake frames." It drastically improved gaming fluidity and smoothness in our playtesting in a handful of games made available to press before launch. AI upscaling is killing native rendering now, and it makes sense for Nvidia to prioritize it in a generation where AMD's Radeon crew already admitted that it isn't targeting highest-end performance. If the DLSS 4 dream fails to materialize, however - if the rest of the 75 launch games exhibit visual artifacts or glitches more pronounced than we've seen, if gamers reject the extra latency AI frames require, a million other ifs - the RTX 50-series could wind up being one of the most disappointing GPU generations in recent memory. We've already seen CUDA core counts correlate to the RTX 5090 and 5080's performance. That's a bad omen for the RTX 5070 series, despite beings priced $50 lower than before. The same CUDA comparisons project the $750 RTX 5070 Ti as being the "star" of this generation, with a roughly 17 percent increase over its predecessor, plus an upgrade to memory capacity, bandwidth, and speed that could help boost performance even more. A ~20 percent performance increase generation-to-generation is still pretty mediocre as far as a progress goes, but it's a more tangible improvement. The reception for the $550 RTX 5070 could be even frostier if our estimates continue to hold; that GPU only has 4 percent more cores than the existing 4070. It sticks to the same 12GB of memory as its predecessor, too, albeit of the faster GDDR7 variety. We'll have to wait to see how it handles in real life when it launches later this spring, but the 5070 has the potential to make the RTX 5080's minuscule generational gains look comparatively massive. Nvidia hasn't announced lower-end GPUs yet. What we've seen from the RTX 5070 laptop GPU makes us nervous about the RTX 5060, as a graphics card with just 8GB of VRAM in 2025 doesn't cut it. We can't opine too much about that until the card is announced and tested though. What's clear is that Blackwell's traditional rendering performance doesn't move the needle forward by much. The RTX 50-series was built for an AI-augmented future and it will live or die based on DLSS 4's reception. Good news! I absolutely adore DLSS 4's Multi Frame Gen in practice. So do Adam Patrick Murray and Will Smith, my cohorts on PCWorld's video team who have been knee-deep testing RTX 50-series GPUs for weeks now. Check out our long discussion about DLSS 4 in the time-stamped video below if you want to hear our deeper thoughts. But make no mistake: Nvidia's entire RTX 50-series lineup bets that DLSS 4's humongous, AI-powered gains in fluidness and frame rates will be embraced by PC gamers. And that is indeed a gamble. DLSS 4 blew my hair back during testing, but many PC gamers derisively call AI-generated frames "fake frames." The RTX 40-series' Single Frame Generation faced withering criticism at launch due to visual glitches and latency concerns. I've seen many, many fewer of those with my hands-on time with Multi Frame Generation - or perhaps they've simply blown by too fast for my eyes to notice, churning out at insanely high frame rates. We've only had access to a handful of MFG-equipped games, and last generation, DLSS 3's Single Frame Gen could be very hit or miss depending on the game you're playing. (Microsoft Flight Sim notoriously had issues with glitchy text, for example). Making DLSS 4 available in 75 games and apps at launch will put a lot of software in gamers' hands for testing, which is excellent and a huge improvement from past norms, but it remains to be seen if they all deliver the same finger-licking performance as Cyberpunk 2077, Star Wars Outlaws, and Alan Wake 2. How to enable DLSS 4 in all those games won't necessarily be obvious, either. While some games are being updated for DLSS 4 by developers, many of the 75 games and apps will require you to enable "DLSS Override" options deep in the Nvidia App, which tells the software to replace older DLSS 3 files with the new DLSS 4 capabilities. Will gamers know to do that? Will those DLSS Override games perform as smoothly as native titles? Will lower-tier cards like the RTX 5070 and 5060 deliver enough raw performance to hit 60 or more frames per second, the baseline needed to mitigate MFG causing devastating, swimmy-feeling lag? Will developers rush to add DLSS 4 to even more titles? Nvidia clearly is gambling on "yes" to all of the above. From what I've played, DLSS 4 can deliver heavenly, snappy gaming experiences that are simply impossible with traditional rendering. But if Nvidia can't convince gamers that AI-generated frames are more than simply "fake frames," the entire GeForce RTX 50-series could wind up being viewed as a flop - because as we've already seen from the RTX 5080 and 5090, Nvidia bet this generation's future on AI and DLSS, not traditional rendering.

DLSS 4 is incredible, but I'm still worried about its future

ReSpec This story is part of Jacob Roach's ReSpec series, covering the world of PC gaming and hardware. Updated less than 4 minutes ago Table of Contents Table of Contents Transforming the DLSS model One frame, two frame The flagship dilemma It's hard to say if the RTX 5090 or DLSS 4 won Nvidia more limelight over the past week. The RTX 5090 is undoubtedly the best graphics card you can buy, but a big part of that conversation has involved DLSS 4 and its multi-frame generation capabilities. And for good reason. DLSS 4 is remarkable, enabling gaming experiences like full path tracing in Alan Wake 2 at triple-digit frame rates without breaking a sweat. Still, I'm worried. DLSS 4 represents a radical change for all RTX graphics cards, from multi-frame generation for the latest RTX 50-series GPUs to a new AI model that reaches down to cards like the RTX 2060 Super. For as impressive as DLSS 4 is, however, I'm concerned that the floor will fall out from under it eventually. Get your weekly teardown of the tech behind PC gaming ReSpec Subscribe Check your inbox! Privacy Policy Transforming the DLSS model DLSS 4 is a much more significant update than you might believe. The headlining feature is DLSS Multi-Frame Generation or MFG, but this new version has tweaked every aspect of DLSS. That means that, yes, DLSS 4 will retroactively make DLSS better, even for older GPUs that don't support MFG. The big change is that Nvidia went from a convolutional neural network, or CNN, to a transformer model. Some games, such as Cyberpunk 2077 and Alan Wake 2, let you switch between the different models within the game. However, I suspect most games with DLSS will just stick with the transformer model moving forward. If, for some reason, you prefer the CNN, you can revert to it through the Nvidia app with its new DLSS override feature. DLSS 4 | New Multi Frame Gen & Everything Enhanced Nvidia says that the new transformer model has more parameters, but more importantly, transformer models are self-referential. Instead of tracking every pixel throughout the scene, the new DLSS model can track specific pixels to improve image quality in problematic areas. Nvidia has demonstrated this capability, which you can see in the video above. There are clear improvements, but they aren't universal. Cyberpunk 2077 - DLSS Transformer Model Above, you can see a comparison between the CNN and transformer model in Cyberpunk 2077. Without labels -- or even with them, actually -- it's basically impossible to tell the difference between the two images. I searched and searched, peeping at every pixel, to find some sort of clear difference between the two, and it's just not there. Both look great, especially considering I'm running in DLSS' Performance mode, but I wouldn't say the transformer model is clearly better. Alan Wake 2 - DLSS Transformer Model The situation is identical in Alan Wake 2, where, once again, it's basically impossible to tell the difference between the two images. There's a slight difference in the detail on the bench, but I'm struggling to attribute that to DLSS when there's some layer of water blurring the camera. But here's the thing -- the new transformer model doesn't need to be better, especially right now. What's important is that it's just as good as the CNN model, at minimum, and with no hit to performance. You'll see the benefits in some games and situations, but the transformer model isn't a glossy new coat of paint that covers all of DLSS. Treat it like one of Nvidia's smaller DLSS updates, which have quietly made the feature better over the past several years. Cyberpunk 2077 - DLSS Transformer Model Scene 2 There are some improvements here, too. Take a look at another scene from Cyberpunk 2077 above. The new transformer model not only reproduces the reflections more clearly, it also deals with a ton of ghosting in the CNN model. There are these long, obvious trails behind each tail light as the cars pass the camera with the CNN model, and they completely disappear with the transformer model. This will likely improve over time. Nvidia has said that it has a supercomputer that's been improving DLSS for the past six years, and that work isn't stopping with the new transformer model. Although the benefits might not show up everywhere, they're still a welcome addition. The transformer model is available to all RTX GPUs, and it covers not only DLSS Super Resolution but also Ray Reconstruction. One frame, two frame OK, but I know why you're all here. Let's talk about MFG. Exclusive to RTX 50-series GPUs like the RTX 5090, DLSS 4 unlocks up to 4X frame generation -- one rendered frame alongside three generated ones. There are 75 confirmed games that'll support DLSS 4, though not all natively. Some will have MFG options in the graphics menu, while the rest will need a DLSS Override from the Nvidia app. And MFG is really good in DLSS 4. The chart above speaks for itself on that front. In Alan Wake 2, I went from less than 30 frames per second (fps) on the RTX 5090 to well over 200 fps. That's an exponential increase in performance and with excellent image quality in tow. Latency isn't an issue, either. Nvidia does a wonderful job these days of limiting the latency hit with frame generation, and adding a couple of extra frames into the mix doesn't significantly increase the latency -- it barely moves the needle. The situation is the same in Cyberpunk 2077. There's barely any additional latency with frame generation turned on, even up to 4X mode. Add DLSS Super Resolution into the mix to increase the base frame rate -- and reduce the total latency -- and you're sitting near 240 fps at 4K. The superpower of MFG is that it's multiplicative; if you feed it with a higher base frame rate, you'll get a larger performance boost out of it. You'll get the best experience out of MFG if you feed it with at least 60 fps, and Marvel Rivals is proof of that. In this title, my base frame rate is acceptable, and latency isn't a concern. In a game like this, I don't need to resort to DLSS Super Resolution. I can just use MFG to improve the smoothness of the game with a minuscule hit to the overall system latency. The overall latency only really becomes an issue if it's high to begin with; as you barrel toward 0ms, even "large" spikes or dips in latency only represent a few milliseconds. If you have a higher base frame rate, even a 50% jump in latency might only move you from 20ms to 30ms. With a lower base frame rate, that same jump might move you from 100ms to 150ms. Same percentage, but the experience couldn't be more different. Marvel Rivals is a good example of this dynamic. However, it also exposes my concern with DLSS 4 moving forward, particularly when it comes to weaker GPUs in Nvidia's lineup. The flagship dilemma Above, you can see a video of Cyberpunk 2077 with DLSS MFG running in its 4X mode. I'm not using DLSS Super Resolution here, so I'm feeding the AI model with a base frame rate of just under 30 fps -- somewhere around 27 fps if I remember correctly. This isn't how you'd want to play Cyberpunk 2077. There's a constant motion blur over everything, along with a ton of visual artifacts. I'm using the RTX 5090 here at 4K, and thankfully you don't need to settle for this experience. Flick on DLSS Super Resolution to increase your base frame rate, and you're off to the races with a smooth experience free of visual artifacts. But we can't all buy an RTX 5090, can we? Something that's important to remember about DLSS is that it's at its best when it's pushing the next level of visual quality; it's at its worst when it's making up for hardware that can't hit the mark. DLSS Super Resolution looks and performs best at 4K. At 1080p, the performance gains are smaller and the visual quality is worse. DLSS Frame Generation is spectacular if you're trying to climb into the triple digits when you already have a playable base frame rate. It falls apart with visual artifacts and unmanageable latency if you don't walk in the door with that playable frame rate. The same problems are here with MFG, and they're magnified. Eventually, Nvidia will move onto RTX 50-series GPUs lower in its product stack. It hasn't officially revealed its budget offerings, but I'd be shocked if we didn't see an RTX 5060 at some point. It's sure to be a popular GPU, likely due in no small part to DLSS 4. But will the GPU itself be powerful enough to capitalize on DLSS 4? Will MFG even be an option for players who can't get to that 60 fps mark in demanding games? The idea brings me back to the Surface Laptop Studio 2. When the laptop was announced with an RTX 4060, I wrote about how Nvidia is enabling a new tier of gaming experiences on hardware that otherwise wouldn't be suited for prestige titles like Cyberpunk 2077. You may get some visual artifacts, and you may not get perfect upscaling. But it's at least possible to play these games thanks to Nvidia's dedication to shortcuts on the rendering pipeline. That dynamic is thrown off when looking at MFG. It can provide what looks like a high frame rate, but without enough power to back that up, it falls apart. The video above is a clear example of that. My concern is that the bottom of DLSS 4 will fall out from under it when weaker hardware is brought into the mix. You can compromise on upscaling; you can even compromise on frame generation in titles that aren't latency-sensitive. Here, with MFG, it's hard to accept the compromises if you can't feed the AI with at least 60 fps. This isn't a critique of Nvidia. I don't believe a $300 graphics card -- or however much the RTX 5060 ends up being -- should match the experience of a $2,000 graphics card like the RTX 5090. However, it is an important note. It's easy to look at the benchmarks and dream up what DLSS 4 could do for budget GPUs, but I'm holding my breath until those GPUs are here to see if they're powerful enough to even take advantage of the feature. That's something that will need to be settled once those GPUs are here, though. For now, DLSS 4 is very impressive. If you feed it with at least 60 fps, it feels like wizardry. Add on top of that a new transformer model that promises even larger image quality improvements over time, and it's clear that Nvidia wants DLSS to remain at the forefront of PC gaming tech.

Nvidia's DLSS 4 is so much more than just 'fake frames'

Here's what you need to know about DLSS 4 and Multi Frame Generation, a key addition to Nvidia's GeForce RTX 50-series. This year at CES, Nvidia presented the next generation of its DLSS upscaling technology, which is trained with the help of artificial intelligence, alongside the new GeForce RTX 5090, 5080, and 5070 (Ti) graphics cards. The company touted its major advantages -- and now that RTX 5090 reviews are live, we can confirm that DLSS 4 indeed feels like black magic, supercharging frame rates and making games feel just as snappy as the beloved Doom 2016. That's because DLSS 4 now supports Multi Frame Generation (MFG), an AI-based multiple intermediate frame calculation that can artificially generate up to three images and insert them between two "real" frames, thus quadrupling the frame rate. Of course, this feature only works on new Blackwell-based RTX 50-series GPUs. But are the AI frames generated in this way a step forward or is it all hogwash? Let's take a close look at DLSS 4 and its multi-frame generation works along with some early impressions. Nvidia DLSS 4 builds atop the existing DLSS 3 and DLSS 3.5 feature set. It's made up of the following five functions: The classic upscaling technology DLSS Super Resolution renders games a lower internal resolution, then upscales to a higher output resolution with the help of AI. Super Resolution also works on older graphics cards from the GeForce RTX 20-, 30-, and 40-series, but it gets more complicated with the other DLSS features. While the simple DLSS Frame Generation (FG) is reserved for the GeForce RTX 40- and 50-series graphics cards, the latest DLSS Multi Frame Generation feature is only supported by the latest generation. DLSS Ray Reconstruction, which improves visual fidelity in ray traced games, is mastered by all RTX graphics processors. Nvidia itself graphically summarized the entire feature set of DLSS 4 by generation. Together with DLSS 4, Nvidia is also introducing a new Transformer AI model, which replaces the previously used Convolutional Neural Network ("CNN") as the neural network behind DLSS. Nvidia promises gamers even better upscaling, more precise ray reconstruction, refined AI edge smoothing (DLAA), and more performance with the switch. We can confirm it looks great in action. The advanced architecture of the new neural network uses the principle of "Deep Attention" as well as the global context to generate significantly sharper details and significantly reduce artifacts such as ghosting. Nvidia DLSS 4 now supports Multi Frame Generation, an intermediate image calculation that can artificially generate up to three AI-calculated images and insert them between two classically rendered images. Standard DLSS 3 Frame Generation only inserts a single AI frame between rendered images. Nvidia shows the difference in two easy-to-understand diagrams below: The new technology uses several neural networks, as well as the Nvidia Optical Flow Accelerator and the optimized tensor computing units inside the new Blackwell GPU architecture powering the RTX 50-series. Note that Multi Frame Generation will not be backwards compatible with prior RTX generations. Nvidia promises huge leaps in performance through the use of DLSS 4 with Multi Frame Generation and makes this clear using the example of the particularly performance-hungry action role-playing game Cyberpunk 2077. In PCWorld's testing, both visual smoothness and responsiveness skyrocketed after flipping on DLSS 4's Multi Frame Generation in Cyberpunk 2077 (using the RT Overdrive preset and 1.7x DLSS scaling). Without any sort of Frame Generation on, the game runs at 71fps; enabling DLSS 4 MFG lets it run an absolutely staggering 251 percent faster, and a just-as-insane 91 percent faster than the 5090 with DLSS 3's single Frame Gen active. Hitting such high speeds looks and feels so good. But what effect do these so-called "fake frames" have on the gaming experience and image quality? Gamers who use Multi Frame Generation must be aware that when it's on, "artificial" frames are added to each "real" rendered frame. But is this ultimately real progress, as Nvidia CEO Jensen Huang described the feature at his contentious CES 2025 keynote, or is it just "fake frames" as some enthusiasts have been shouting from the rooftops in online forums? Of course, the whole thing needs to be considered in a differentiated way and in independent tests. It's understandable that buyers of a $1,999 graphics card want significantly better performance without "tricks" compared to the previous generation. (Fortunately, the RTX 5090 delivers solid traditional performance gains too.) But as the calendar flipped over to 2025, PCWorld declared that AI upscaling killed native graphics gaming. We're better off for it. After using it, DLSS 4 feels like it drives that theory home. Whether DLSS 4 adds true extra "frames" in the traditional sense or acts more like motion smoothing on steroids is an important technical distinction that demands further discussion and testing. But there's no doubt that games look and feel so much smoother and snappier with DLSS 4 MFG and Nvidia Reflex active. It provides a big boost to the quality of your gaming experiences, period (though I'd still hesitate to turn it on in competitive multiplayer games). We have more DLSS 4 coverage coming, but here's a tease from our RTX 5090 review: Turning on DLSS 4 makes Star Wars Outlaws, a fun game prone to performance concerns, feel just as good as the legendary Doom 2016, which many gamers consider the paragon of fast-action shooters. "It's like a whole new game," PCWorld's Will Smith says. It's high praise coming from a guy who has been reviewing graphics cards and making games for decades. The use of artificial intelligence and features based on it (like DLSS 4) should be welcomed if, for example, intelligent AI upscaling improves both the picture quality and the frame rate -- like DLSS Super Resolution does. The intelligent and extremely powerful AI edge smoothing using DLAA is also a gain, especially for enthusiasts. The fact that Nvidia no longer compares its new generation of graphics cards with native rendering performance from the previous generation, instead marketing somewhat "embellished" frame rate values with long bar charts in benchmarks, isn't very helpful. Comparisons of a GeForce RTX 4090 with DLSS 3 and Frame Generation against a GeForce RTX 5090 with DLSS 4 and Multi Frame Generation say absolutely nothing about the actual increase in rendering performance and are a source of uncertainty for buyers. The poor initial reception by gamers is understandable and justified. That being said, now that we've tested it, DLSS 4 truly is a revolutionary, game-changing technology. Trying to show such a gigantic experiential leap forward is hard to do with simple frame rate graphs. I don't like that Nvidia failed to show how the RTX 5090 compares to the 4090 in traditional performance during its blockbuster CES keynote, but it's somewhat understandable now that we've experienced the difference with our own eyes and hands. DLSS 4 is that good. Perhaps the bigger takeaway here? Hardware and software-side AI features such as upscaling, frame generation, latency improvers, and dedicated AI hardware are here to stay and will play an even greater role in the future. The days of native rendering are coming to an end. You certainly don't have to like that, but I think enthusiasts need to start accepting it. The game has changed.

Should you play your games with Nvidia's DLSS 4 on or off?

Image of Black State, showcasing the light bulb transitioning from translucent to transparent when DLSS is turned off and on. (Image credit: Motion Blur) Nvidia's new RTX 50-series GPUs are making waves across the gaming world, but what's more interesting than their power are their features, specifically DLSS 4. The company is making some bold promises about how much DLSS 4 -- the fourth generation of Nvidia's deep learning super sampling technology -- can use AI to increase performance and enhance visuals seemingly instantly. There's some magic behind the scenes that goes beyond it's powerful -- to paraphrase marketing language we often hear about such tech. But if you've read any fantasy novel, you know that magic comes at a cost (or it should, and if it doesn't, that's bad writing). So what's the cost of having DLSS 4 casting its spells? I chatted with an Nvidia rep and the Studio Director of Black State to get some answers. Let's dive in. My first worry was about art style and art design. Nvidia shared images that showcased how DLSS 4 completely transforms scenes, objects, and character models. Some transformations are so dramatic that the art of the game itself feels different. I'm thinking of the light bulbs in Black State and the hair in Indiana Jones and The Great Circle pictured above. I asked an Nvidia spokesperson about what DLSS 4's impact will have on the art design, both retroactively and in games currently in development. "Developers continue to have full control, as DLSS requires game integration," the rep tells Laptop Mag. This includes when DLSS is added retroactively. DLSS is trained from synthetic data to accurately reconstruct the final image based on the ground truth image of the game as envisioned by the developers." In other words, the implementation of DLSS 4 is in the hands of the developers, so whatever you see when DLSS 4 is enabled is, technically, how the game was intended to look. One of the examples that the Nvidia spokesperson pointed toward was Ray Reconstruction, which makes lighting more stable. An image of Cyberpunk 2077 showcased how an in-game advertisement was perfectly reflected in a puddle on the ground (compared with it off, which showed a muddier version). However, this means developers are also responsible for how the game presents itself when DLSS 4 is disabled. With features like RTX Hair, Ray Reconstruction, Multi Frame Generation, and more, I hope developers take care to ensure gamers who don't own the latest tech can still experience the game with the same impact. That's exactly what devs are doing, according to Ömer Faruk Güngör, Studio Director of Motion Blur, the developer of Black State. "We constantly monitor everything we add to [Black State] to ensure it provides the best visual quality and performance for all configurations, not just the latest ones," tells Laptop Mag. Güngör went on to say that there will be differences when DLSS 4 is on or off, but Motion Blur is "building the game considering both options should look stunning and be well-optimized." But if games are being built with DLSS 4 in mind, especially considering Multi Frame Generation, does that encourage developers to slack off when considering the native art? I mean, why optimize your title when AI can do it for you? Well, in the case of Black State, Motion Blur is "optimizing the [game] as if native resolutions are the only options available." As far as the balance of art design and technology, Motion Blur is "working meticulously to provide an incredible experience for all gamers." This is the ideal response when new technology arises. AI can be a useful tool, but not at the expense of jobs and originality. Outside of the core changes to the art, Nvidia boasts that the performance DLSS 4 will bring is better, better, and better. The Nvidia representative I spoke with said, "DLSS provides higher frame rates, lower latency, and better image quality than native." They went on to say that, "For developers, the extra performance headroom from DLSS can mean achieving higher-quality visuals with more detailed textures, more realistic lighting, and more geometry, without compromising on performance." Let's get real: We are in a technological age where bigger isn't always better. In terms of raw performance, higher fps means it's better, yes. But DLSS 4 introduces Multi Frame Generation, which produces AI-generated images to increase the number of frames on screen. There are 2x, 3x, 4x modes. What that indicates is how many AI-generated frames are in between true frames. At Frame-Gen 2x, there's one AI-generated frame for every true frame. For Frame-Gen 3x, it's two, and for Frame-Gen 4x, it's three. The generated frames are informed by the previous true frame and the following true frame. Then the AI fills in the gaps. However, that kind of processing is sure to affect the quality, latency, and fluidity of the game, right? I followed up with Nvidia about the impact of AI processing on gaming quality but haven't heard back just yet. When I do, I'll update this article. Despite that, Digital Foundry got an early access look at an RTX 5080 GPU running with Super Resolution and Multi Frame Generation in Cyberpunk 2077. What we saw was a massive increase in frames. They could not provide real numbers, but the percentages jumped from 100% (off) > 563% (2x) > 749% (3x) > 956% (4x). But what about latency? Well, that's complicated. We didn't get to see it when it was off, but a trip through the Night Market revealed the times for Frame-Gen 2x (50.97ms), 3x (55.50ms), and 4x (57.30ms). Given the increase in frames, that trade-off doesn't seem bad at all. However, that's what we see on a compressed YouTube video. We actually won't know what the trade-off feels like until we're playing these games on our own monitors. Now, the biggest catch here is the visual artifacts. As we all know, AI generation isn't perfect. To showcase how the GPU handles 120 fps, Digital Foundry had to slow the video speed to 50% (because YouTube caps at 60 fps). That means the visual artifacts you see in the video will look worse than actually playing the game (probably). Some people might not be bothered by it, but you can't deny that when the camera passes by neon signs or trees, aspects of it look like pixelated garbage. That may seem harsh, but that's just what visual artifacts are. The question is, are developers keeping these issues in mind when creating new games? According to Güngör, "As a tech-oriented studio focused on delivering the best quality visuals with the best performance, we closely monitor the impact of every addition to the game in terms of milliseconds. With the introduction of DLSS 4, we have more room for enhancing quality." Güngör goes on to say that Motion Blur has considered upscaling technology from day one, and it is "prioritizing technological advantages for all gamers to ensure the best possible performance." However, it isn't always in the developer's hands. One interesting feature that comes with DLSS 4 is DLSS Override. This can override games integrated with legacy DLSS models in favor of DLSS 4. I have some serious concerns about this, as this definitely falls out of the "developers have full control" category. We won't be able to see the impact of this until we test it, but you cannot deny that the game will be presented in a way outside of the developer's intention, even if it looks better. While we haven't gotten a close look at DLSS 4 yet, developers are working hard to improve it alongside Nvidia. Güngör states, "The DLSS technology is evolving into something better every day. We are working with the Nvidia team closely not just implementing it but also giving feedback for making it better." I can talk about the potential good and bad all day, but right now, it's all wait and see. We'll start seeing RTX 50-series gaming laptops launch in March and April, so keep an eye out for our reviews and best gaming laptops page for these new devices to fill in the slots.

Nvidia's DLSS 4 just came to your old GPU -- here's what you can use

Table of Contents Table of Contents DLSS Multi Frame Generation DLSS 4 Super Resolution What games use DLSS 4? Get more frames without DLSS 4 Nvidia's Deep Learning Super Sampling just hit version 4.0 and it's the most impressive version yet. The flagship feature is multi frame generation, which made a big splash with the debut of the RTX 5090 -- along with grandiose performance claims -- but it's not the only one. DLSS 4 also introduces other neural rendering features, better upscaling, and a major overhaul of how all of its RTX features are handled, using a new real-time "transformer," which uses the same kind of architecture as modern large language model AIs, like ChatGPT. All of this makes DLSS 4 more powerful than ever before and it's available across all RTX graphics cards -- albeit with frame generation locked to the last couple of generations of Nvidia cards, at least for now. Recommended Videos DLSS Multi Frame Generation The landmark feature of the latest DLSS is Nvidia's Multi-Frame Generation, an AI-powered technology that can generate up to three additional frames using AI, based on a single rendered frame from the GPU. This is what allowed Nvidia to launch its RTX 5090 with such monstrous claims of "double RTX 4090" performance, and achieve over 200 frames per second (fps) in games like Cyberpunk 2077 and Black Myth Wukong, even at 4K with ray tracing enabled. There are some caveats to Multi Frame Generation, namely that you need at least 60 fps (even if using DLSS upscaling) to avoid some of the worst artifacts that can come from the AI trying to generate frames without enough information. There can also be a strange ghosting, or motion-blur-like effect at times, so the 4X frame generation setting is one that will probably see less usage than the 3X or even 2X, which can both provide a sizeable uplift in frames per second. Frame generation and multi frame generation in particular, also introduce additional input latency, but Nvidia's new Reflex 2 technology does a good job of balancing that out. Unfortunately, for now, DLSS multi frame generation is an exclusive feature of the RTX 50-series, like the RTX 5090 and 5080. There's some talk of it coming to older graphics cards like the RTX 40-series, and even RTX 30-series with optimizations, but that's very much up in the air. DLSS 4 Super Resolution The core feature of DLSS is its Super Resolution upscaling that lets the game render at a lower resolution to save GPU power, and then has AI upscale the frames to appear higher resolution. It comes in a range of quality modes depending on user preferences, and there's also the option of Deep Learning Anti-Aliasing which just upscales for added quality, without enhancing performance. It also works with every Nvidia RTX graphics card, from the RTX 20-series right through to the latest RTX 50-series. Marvel Rivals - DLSS 4 Gameplay Whatever Nvidia RTX GPU you have, you'll be able to take advantage of DLSS 4's enhancements to DLSS Super Resolution. The new transformer model enhances DLSS image quality dramatically over the previous Convulitional Neural Networks (CNN) design, allowing for an increase in the number of sampled parameters from the rendered frame, giving DLSS more information to work with in upscaling the image. That in turn reduces ghosting and visual artifacts, improves motion resolution, and smooths edges far better than before. This transformer model also improves the effect of additional RTX technologies like Ray Reconstruction, making ray tracing more effective on finer objects with viewer visual artifacts and less noise. Better yet, this is just the first generation of Nvidia's transformer model. It spent six years refining its CNN model, so we should see continued DLSS image quality improvements in the years to come, that can be enjoyed by all Nvidia RTX card owners, regardless of generation. What games use DLSS 4? One of the biggest flaws of previous DLSS generations is their lack of early support. When the first version of DLSS launched with the RTX 20-series cards in 2018, it was so new and fresh that only a handful of games made us of it. In the years since, we've seen hundreds of game developers adopt DLSS technology, alongside competing standards from AMD and Intel: FSR and XeSS. With DLSS 4, though, Nvidia hit the ground running. On Day 1 of its RTX 5090 being available, it had over 75 games supporting DLSS 4 ready to play. You can find the full list here, but standouts include Indiana Jones and the Great Circle, STALKER 2, Alan Wake 2, Black Myth Wukong, Cyberpunk 2077, and more. In games without native support, Nvidia is now offering DLSS overrides through the Nvidia app. This will allow you to force the transformer model, as well as force quality settings like Ultra Performance and DLAA that aren't natively included in every DLSS game. Get more frames without DLSS 4 If you aren't set to buy an RTX 50-series graphics card or don't have an Nvidia GPU at all, you won't be able to enjoy DLSS 4, or any version of DLSS for that matter. You can still make use of FSR and XeSS in compatible games for some upscale-enhanced frame rates, but what about frame generation? While Nvidia's "RTX" version of that remains locked to its own hardware, you can use Lossless Scaling to enjoy frame generation in almost any game. It's a $7 app on Steam and we're big fans. Check it out if you're getting some FOMO from all the new RTX 50 news.

DLSS 4 Explained: Everything You Need to Know About Nvidia's Latest AI Upscaling Tech - IGN

Nvidia has been a market leader in PC gaming graphics for years, and with the rise of machine learning, AI is playing an increasingly important role in how we experience our games. One of its most revolutionary advancements has been Deep Learning Super Sampling (DLSS), an intelligent upscaling solution that opens the door to increased performance, especially at higher resolutions. This technology has undergone multiple iterations over the years, and with the launch of the RTX 50-series, it's making its biggest jump forward yet with DLSS 4. In this guide, I'll explain everything you need to know about how this tech works, why it matters to you and the future of PC gaming, and why it's something you may want to keep in mind with your next PC upgrade. DLSS, or Deep Learning Super Sampling, is Nvidia's proprietary system for intelligently upscaling games. The company has consistently developed and built upon DLSS since its debut in 2019. Throughout that time, its core purpose has been to improve performance by rendering games at a lower resolution and then upscaling that content to your monitor's native resolution. Rather than leave you with the soft and potentially blurry image you would experience turning up the resolution yourself, DLSS applies its scaling through a neural network that has been trained on thousands of hours of video games. Alternatively, if you don't want or need to upscale, you can instead enable Deep Learning Anti-Aliasing (DLAA), which provides image enhancements to your native resolution. These features are only available on Nvidia graphics cards with Tensor Cores, which began with the RTX 20-series. This is because the upscaling and enhancements taking place are the result of thousands of hours of neural training on Nvidia supercomputers. Its neural network ingests and learns from huge data sets to learn how to upscale and reconstruct/enhance images with the least quality loss and, in fact, even provide additional clarity in some situations. As time has gone on, Nvidia has enhanced the system with additional features. One of the biggest is Frame Generation, which uses artificial intelligence to create an additional frame between each rendered frame, increasing frame rate. When used in conjunction with Nvidia Reflex (which is also being enhanced with the 50-series), these additional frames can blend in with a minimal impact on latency. The result of which isn't just "better performance," but allowing lower-performance graphics cards to reach previously unattainable frame rates with aspirational graphics settings. Which brings us to the current day with DLSS 4. This generation marks the most significant change in the feature's history and includes a completely different, and much more capable, AI model. The Transformer Model So far, DLSS has used a model of AI known as a CNN, or Convolutional Neural Network, to deliver its benefits. This type of model analyzes an image to determine key elements, like lines, edges, and spatial relationships to determine how to apply its enhancements. It's a very common type of neural network specializing in image analysis, which is why it's not surprising that it was the foundational model behind DLSS up to this point. DLSS 4 instead utilizes a Transformer model. A Transformer is a different form of AI model that's able to calculate twice the number of parameters to better understand each frame of the scene. Put another way, it's able to understand what it's looking at and what's taking place much better, and then apply more sophisticated calculations to deliver a higher quality image. This new model is core to DLSS 4 and impacts each of the different pieces that allow DLSS to look and perform so much better than the prior version. With DLSS 4, Deep Learning Super Sampling is much more than a simple upscaler (and never was, really). Instead, it's a series of systems that work together to improve performance, enhance image quality, and reduce latency. In addition, frame generation has been significantly upgraded and can produce three times as many frames as DLSS 3.5. It's this network of cooperative systems that allow DLSS to look and perform better than it has in the past. One system handles upscaling (DLSS Super Resolution), while another handles lighting and shadows (DLSS Ray Reconstruction). DLSS Frame Generation multiplies the frames, while Reflex 2.0, which is it's own feature separate from DLSS, keeps latency numbers low so you're not noticing input lag as you play your games. DLSS Super Resolution is the sub-system that handles upscaling. If you've ever noticed DLSS in an in-game menu, there's a good chance that there were quality options to choose from: Ultra Performance, Performance, Balanced, and Quality are the typical presets. Each of these levels adjusts the game's rendering resolution, the base resolution that the neural model will need to upscale. Through the new TNN model, DLSS Super Resolution is able to deliver much sharper results and retain much more detail that the previous CNN model would lose, including in motion. The results can look nearly as, equal to, or even slightly more crisp than native resolution. While any enhancements beyond native resolution are subject to discussion and scrutiny, when it works, it can look fantastic. The impact of the TNN is especially noticeable in fine detail, like textures, fine edges, and lettering. DLSS Ray Reconstruction is the second core image enhancer and one that has seen big improvements. Replacing traditional denoisers (systems that remove graininess and "noise" from a scene), this portion of DLSS is focused explicitly on analyzing and reconstructing lighting and shadow information. Like DLSS more broadly, it has been trained on thousands of hours of data to improve its understanding of different lighting conditions and how they should appear in real-time renders. The TNN model provides DLSS Ray Reconstruction with much better understanding than it was previously able to have, and the results are easy to spot. DLSS Ray Reconstruction in the CNN model often struggled with fine lines and moving shadows. In the image above, the results speak for themselves, but it's also evident elsewhere. The flickering that was previously evident in distant shadows and lines (like telephone wires) is dramatically reduced. Objects in motion, like ceiling fans, retain more clarity. The strange action of "bubbling shadows" isn't nearly as evident. While scenes in games are variable and more testing and iteration is necessary to come to any hard conclusions, it's difficult to argue that it's not a leap ahead for DLSS as a system. Deep Learning Anti-Aliasing, or DLAA, is an alternative to DLSS Super Resolution. If you don't need the upscaling features, DLAA allows you to significantly enhance your native resolution through TNN-enhanced anti-aliasing. DLAA smooths out edges much better than traditional anti-aliasing models and maintains those improvements in motion. The result is a very sharp image that looks noticeably clearer than gaming at native resolution with standard anti-aliasing. For gamers with older GPUs, you'll be able to upgrade to the new Transformer model, as well as toggle DLAA or DLSS Ultra Performance mode, within the Nvidia App. DLSS Frame Generation made its debut with the RTX 40-series. It was contentious at the time, though has largely been accepted as an effective way to improve your in-game performance. This system, which I'll refer to as single frame generation (SFG), allowed the GPU to leverage its Tensor Cores to create an artificial frame based on the details of the previous frame. Using this technology, gamers could play at higher resolutions and frame rates than may otherwise have been possible, make better use of their high refresh rate gaming monitors, and enjoy smoother action overall. With DLSS 4 and the new Transformer model, the system is now able to generate up to three artificial frames for every true frame that's rendered. This new capability is DLSS Multi Frame Generation (MFG). DLSS is able to accomplish this thanks to the improved performance capabilities of the new TNN, as well as shifting optical flow into a neural network instead of relying on the hardware-based Optical Flow Accelerator on the RTX 40-series. Because this AI Optical Flow system is unique to the RTX 50-series, DLSS Multi Frame Generation is exclusive to this generation for now. Optical Flow, in general terms, is the AI's ability to interpret the composition and motion within a scene to determine what should be rendered into its neural frame. Because the transformer is able to analyze each scene more thoroughly, ingesting more data points, it's able to more accurately anticipate what will occur further into the future. While it would be easy for things to get messy with the TNN rendering 75% of the frames when set to its max, the RTX 50-series also introduces flip metering. The important thing to know here is that flip metering controls frame pacing to ensure that gameplay remains smooth. However, Nvidia does recommend that MFG only be set as high as it takes to reach your monitor's refresh rate. Overshooting to get the highest FPS possible can introduce visual artifacts due to the mismatch. DLSS 4 is only one piece of the AI-enhanced PC gaming future Nvidia has promised us, but it's an exciting one. While it's clearly designed to have something for everyone, its benefits are likely to be especially keen for gamers playing on mid- to low-performance GPUs, extending the usable life of the hardware and opening the door to higher resolutions and graphics settings than would otherwise be possible. Time will tell how DLSS 4 ultimately shapes up, and if prior generations are any indication, Nvidia can be expected to build upon this foundation throughout the generation. As a starting point, however, DLSS is poised to impress.