Will new technology hold up against DLSS?
AMD Introduced With The Launch Of The Radeon RX 6X50 XT Series FidelityFX Super Resolution 2.0, AMD FSR 2.0, debuts. This technology has a goal similar to that of Nvidia Deep Learning Super Sampling (Nvidia DLSS), and the latest Intel Xe Super Sampling (Intel XeSS): to make the game possible at a lower resolution, use image augmentation technology while retaining as much detail as possible, thus offering more performance with reduced image quality.
More than just an evolution, FSR 2.0 is a completely different technology than FSR 1.0
AMD FidelityFX Super Sampling 2.0 is AMD’s “built from scratch” technology, so unlike version 1.0, it almost deserves a name of its own. Let’s reveal AMD’s new strategy, the result in performance and graphics quality and also how it feuds with Nvidia DLSS. Do we no longer need strung cores, and thus, RTX, to do a good upgrade?
Really AMD DLSS
The first big change from AMD FSR 2.0 is that it is very similar to Nvidia’s approach to DLSS, or rather DLSS 2.0. AMD FidelityFX Super Sampling 1.0 has done the same with all the technologies listed so far, but with a different technology. He would take an image, enlarge it, and use FidelityFX Constrast Adaptive Sharpening (FidelityFX CAS) to preserve and define image details.
The new approach is closer to DLSS
AMD FSR 2.0 now uses more information, with a time filter. In other words: when determining the shape of each pixel, FSR 2.0 takes into account information from previous frames. It will also use other information from the game engine, and motion vectoror motion vectors, which will tell this technique where things are and where they are headed, again helping to make the right choices for resizing the image.
If you are facing deja vu, you are not wrong, because you most likely remember our article about DLSS 2.0 and how well it works with super sampling technology on the GeForce RTX side.
FSR 2.0 takes motion vector information, object depth information and colors, mixes it with data from previous frames, and decides how to fill in the missing spaces to increase image resolution.
The FSR approach is very similar to the DLSS approach in its modes. Essentially, the image is actually scaled down, and it uses more machine learning to fill in the gaps in higher performance modes. Here is a progression of these modes, remembering that they refer to the final resolution, or output resolution, in this table:
– Continues after announcement –
Comparison of DLSS vs. FSR 2.0 in the video
FSR 1.0 vs. FSR 2.0 and DLSS graphics
It’s time to find out the evolution against FSR 1.0, and what the disagreement is with DLSS in terms of graphics. We’ve done comparisons with 4K resolution and also Full HD resolution, with intermediate entry cards from both Nvidia and AMD. For the first test, let’s see the performance in a 4K scene:
Below we will analyze the evolution of FSR’s performance, balanced modes, and quality in this landscape. To make the difference more visible, we use 5x zoom to capture the details:
Excellent quality setting result. It’s clear how she managed to retain a lot of detail, and thanks to the sharpening effect, we feel that the resolution is even higher than the image already rendered at 1080p. Returning to performance mode, the image begins to deteriorate, especially showing more artifacts on moving objects, something that becomes even more evident if you watch the comparison video. Despite the quality loss, it’s still entirely possible to run FSR 2.0 in performance mode, especially when you’re not with the “nose sticking to the screen” which is the effect of the 5x zoom.
Now let’s go to the comparison again, but put FSR 2.0 to compare with 1.0 and also Nvidia DLSS:
– Continues after announcement –
With 4K performance mode putting the three technologies into practice, i.e. all rendering in Full HD and upscaling to 4K, FSR 2.0’s evolution is remarkable. The image has much more definition, retains more detail and contains fewer artifacts at the edges of moving elements. DLSS still has the advantage, with more defined edges, and fewer artifacts on objects, but the difference is really minimal, especially without the zoom we used in the comparison above.
Let’s move on to another scene, now to explore the differences between these technologies in their quality modes, i.e. from 2560 x 1440 to 3840 x 2106.
We have two magnified sections in this scene:
In this part of the comparison, still images don’t do justice to the more obvious elements of the video. Native 4K with TAA and FSR 1.0 both suffer from serious flickering issues at the edges of objects. FSR 2.0 and DLSS either reduce or correct most of them, providing a much more complete and better picture than the original 4K display. DLSS has the advantage and handles some patterns better, such as grids stuck on the roof of a building, but in this comparison, FSR 2.0 better solves character movement in the background. Here I recommend again watching this scene compared to the video.
Now let’s see what FSR 2.0 technology at lower resolution looks like. In this case, we go with a Radeon RX 6600 to handle the game in Ultra, RT On, and Full HD resolutions.
Again, flickering is an item that will only be visible in the video comparison, but I can already confirm that it was clear how much the FSR 1.0, even in quality mode, suffers from a lot of this problem. FSR 2.0 in Performance mode also had this problem, in addition to a loss of definition to a similar degree as what happens in FSR 1.0 Quality. In short: For 1080p, the ideal is to target the FSR 2.0 quality mode. The graphical score is very good, and while the performance gains are modest, it’s much more interesting than running locally and applying TAA.
This turned out to be an inevitable effect of working at very low resolutions. Like DLSS, FSR 2.0 suffers when the final resolution is too low. In Quality mode, it renders from HD to render Full HD, and in Performance mode we have 890×540 as base resolution, which is little information to rebuild.
Performance of FSR 2.0
We compared the performance of FSR 2.0 with version 1.0 of the technology and also with DLSS, which was implemented in Deathloop in version 2.3 of Nvidia technology.
Entrance and ancient signs
When starting the comparisons with older cards, we see that the popular Radeon RX 580 and GTX 1060 6GB benefit a lot from performance, especially the Radeon. Switching to the latest cards, such as the GTX 1650 GDDR6 and Radeon RX 6500 XT, we get two opposite results.
The GeForce GTX 1650 GDDR6 benefited a lot from FSR 2.0, with gains of 20%, while the Radeon RX 6500 XT practically stuck with a very poor frame rate. In fact, this card did all the tests poorly, implying that the limitations of this model – which we’ve already put on our feet in the past – may have negatively affected this test.
AMD FSR 1.0 still makes the biggest performance gains, but as we’ve shown in the comparisons, it’s not on the same level as the other two technologies.
Starting with Full HD cards, like the RTX 3050 and Radeon RX 6600, we have good results that enable FSR 2.0 quality, and interestingly enough, GeForces grabbed the most benefit. The RTX 3050 increased its performance by about 25%, with a 15% gain that is still good, but much less significant. DLSS performed slightly better than FSR 2.0, while FSR 1.0 and its simpler implementation remains the best test result.
And finally, we got a look at 4K with the RTX 3060, a card that doesn’t have game-handling performance at this resolution. It gave all the tech a respectable boost, with FSR 2.0 already making a 30% gain and setting the average at 50 fps. This shows that simply using FSR in performance mode, which is still a good level of quality, or changing the preset from Ultra to High, to make 4K gaming workable on the RTX 3060. But it’s better to take advantage of the tensor cores and enable DLSS, which will make Gameplay up to 60fps+ in high definition.
AMD FSR 2.0 still has some minor flaws compared to DLSS 2.3, But it won’t even need to hook up or beat Nvidia’s resources. Just close enough To show itself as a valid alternative for Radeon owners. him too.
AMD FSR 2.0 didn’t even need to beat DLSS, just get close enough. and got it
Version 2.0 of FSR is comparable in quality to DLSS, and if it weren’t for our magnified reviews it would be hard to tell the difference. As the gameplay continues, it’s unlikely that anyone will see the difference, unless you stick the pixels under a microscope and stare. This is great news not only for owners of AMD cards, but also for owners of GeForces without DLSS support.
But that’s not all a hack, and AMD FSR 2.0 needs more developer intervention. Fortunately, DLSS 2.3 paved the way for this type of temporal filter through reconstruction, so adoption is likely faster than DLSS in its early days. But it’s still something to see in practice.
Owners of DLSS-capable cards should continue to take advantage of Tensor cores in their GeForce RTXs. DLSS is faster than AMD FSR, offering more frames and slightly better graphics results. After that, AMD FSR is a great option, and 4K should be used even in performance mode, while Full HD is ideal for use in quality mode. If that doesn’t provide enough frames to play, the way is to use FSR 1.0, which is more interesting than not using the feature, in many cases it might be better to stick with the original resolution in those cases.
Source: AMD Community, GPU Open