- Nvidia revealed its latest DLSS 5 generation with advanced neural rendering tech.
- It alters the original art design and causes a massive 50-60% performance loss.
- The version has been rather controversial, as Nvidia has totally lost the plot here.
Nvidia introduced DLSS in 2019 with its RTX 20 series GPUs. The tech didn’t seem perfect at that time, but it had a lot of potential. It is essentially an upscaled image from a lower-quality, allowing for better FPS at higher resolution.
Over time, the tech has become really impressive and, for better or worse, a general requirement for most modern games. Over the years, others like AMD and Intel have also introduced their version called FSR and XeSS.
However, DLSS remains proprietary and arguably the best of all. It retains the best image quality after upscaling from a lower resolution while maintaining higher FPS with next-to-no delay if you don’t use frame generation.
Earlier this year, Nvidia introduced DLSS 5 as its next iteration, and it’s safe to say it’s been quite controversial. Although developers have full control, the tech completely alters the art design, putting kind of an AI filter over things.
Worst of all, Nvidia has revealed that enabling DLSS 5 will cost a whopping 50-60% performance loss, which defeats the whole purpose of the tech, which means Nvidia has, in a way, lost the plot with its upscaling tech.
Why it matters: Since the new version costs massive performance, and even the art style doesn’t sit well with most people, it takes away from the tech’s major selling point.
What Exactly is DLSS 5?
So, the first and foremost thing is to understand what DLSS 5 exactly does. Older versions like DLSS 4 and earlier render the game at a lower resolution and upscale it to a higher resolution using AI to save performance.
The newer versions, like DLSS 4.5, also replace traditional noise-removal filters with AI to clean up ray-traced reflections and lighting. However, DLSS 5 is a lot more complicated and intricate than recent versions.
It uses neural rendering/shading. Instead of just blowing up pixels or cloning frames, the neural network actively creates visual elements that the game engine does not calculate due to real-time performance limits.
This results in a complete alteration of the art style and what the game would originally look like, as we’ve seen from the demos. And while sometimes its genuinly amazing, at times it can feel very off-putting.
But the developers do have complete control over how aggressive it will be, so this issue can be negated. What can’t be ignored is the massive 50-60% performance hit that the tech will cost in its current state.
Where Nvidia Has Gotten it All Wrong
As stated earlier, the main goal for DLSS was to improve FPS by upscaling the image from a lower-quality image to make it look pleasing at the same time. The performance is essential here, which has now been neglected.
Even in the current state, which is expected to be the same at launch, DLSS 5 performs terribly. Nvidia revealed that they were able to boost the performance 5x from the first demo, giving a rough idea of how intensive it is.
What, in my opinion, has the company gotten completely wrong is that no one ever asked for the new neural rendering tech. All everyone wanted was even better image quality while attaining good FPS at the same time.
Changing the whole art style is something that defies what the developers created, and attaining photorealism isn’t always the goal. And even if someone asked for it, no one wants to affect their perfectly smooth frames.
Now, there is no denying how impressive the rendering looks in certain case scenarios, but that is very niche and, let’s be honest, unwanted. What Nvidia could have done was very simple in contrast to this.
What They Should’ve Done
DLSS 4.5 was truly impressive. The earlier versions had certain issues with artifacts and rendering foliage and stuff, but that version mitigated almost every issue, turning Nvidia’s upscaler into an absolute monster.
AMD has come pretty close with FSR, but DLSS is still superior in every way possible. Plus, frame generation also comes in really handy when implemented well by the developers, especially in single-player titles.
To be very honest, fans didn’t really want Nvidia to change much with DLSS 5. All they wanted was even superior quality instead of these artifical looking filter, which no doubt are highly sophisticated and advanced tech.
But if the company worked on improved image quality while retaining the art style, and focused on improving multi-frame gen and reducing its latency, that would’ve been a major win for most Nvidia users.
What are your thoughts on Nvidia’s DLSS 5 fiasco? Let us know your opinions in the comments down below.
Thank you! Please share your positive feedback. 🔋
How could we improve this post? Please Help us. 😔
News Reporter
Abdullah is an avid gamer who primarily plays single-player titles. If you can’t find him anywhere, he’ll probably be at his desk playing The Witcher 3 for the millionth time. When he isn’t playing games, he’s either reading or writing about them.
Join Our Community
