NVIDIA RTX 50 GPUs did not meet the best reception. Between the low availability, the increases in disappointing generational performance and the high prices, it was difficult to justify the expenses for one of the Nvidia Best graphics cards. The saving grace, at least if you believe Nvidia marketing, is DLSS 4 – a software functionality that promises to quadruple your image frequency in supported games.
It is difficult to discuss with a 4x increase in the frequency of images, but even with one of the new NVIDIA graphics cards on my PC, I barely touched the functionality. Apart from playing with DLSS Multi-Frame Generation (MFG) to see how the functionality works, I left it almost with each game I played. Here is why.

Related
I think the arguments around the DLS are exaggerated
Scale technologies like DLS are there to stay, whatever the company you get your GPU. I used the three (FSR, Xess and DLSS) since their release, and although I agree, they were lower than those to start, the improvements in annual shift made it really difficult to identify the difference between a native returned scene and one using DLSS or otherwise. Although I do not use it for competitive games, it is perfect for narrative solo-player experiences like God of War Ragnarök. Does anyone else prefer to play with DLSS, so they get higher image frequencies than the pure raster?
DLSS 4 has a narrow concentration
The high -level concept of quadrupling your image frequency is attractive, but putting these figures in context explains certain problems. Like most PC players, I have a monitor with a cooling rate of more than 100 Hz. For my display, it ends at 138 Hz, but you can replace it with hundreds of game monitors with a refresh rate of 144 Hz and set out with the same conclusion. There are certainly displays with a higher cooling rate, but they are in the minority and are generally intended for competitive players who play 1080p.
In order to completely saturate a 144 Hz screen with DLSS 4, you need a basic image frequency of at least 36 IPS in 4x mode and at least 48 IPS in 3x mode. It’s great in the few games where I sit through these image frequencies, but in the vast majority of games, I am not. In a game like Alan Wake 2, I am barely capable of gathering 20 IPS, while I am in Marvel Rivaux, I can go up to three digits without even thinking about it. Overall, the games fall into one of these two camps, and I should play constantly with graphic parameters to develop them ideal for MFG.
As I am going to search the next, MFG calls for a high basic images frequency to function properly – at least 60 IPS, depending on my experience. Everything that is lower will produce many visual artefacts, especially in 4x mode. I need to use other performance strengthening measures in a game like Alan Wake 2 To even benefit from the MFG – when I don’t even need all additional executives – and I don’t need additional executives for their Marvel Rivaux.
I use these games as touch stones, but you can apply the same concept elsewhere. Black myth: Wukong is too demanding to benefit from the MFG alone, and Galactic deep rock is light enough for me to need any generation of a framework. The same dynamic is true in Silent hill 2 And The finals, Or Star Wars Outlaws And Diablo 4. You have the idea.

Related
The functionality is largely targeted to a few specific use cases. You play a game at 60 IPS and you have a monitor at 240 Hz. Or, you play a game at 90 IPS, and you have a 360 Hz monitor. In these situations, the MFG is excellent as a smooth movement tool that allows you to enjoy the full cooling rate of your screen. The situations where it is really useful are only rare.
For me, not only do I have a monitor with a low refresh rate (relatively) low, but I also do not start a ton of games that fall perfectly in the performance targets where the MFG is useful. In each game that I have played since the upgrade, I was able to achieve the performances that I natively want or thanks to a combination of DLSS super resolution and 2x frame generation.
It collapses when you need it most
Not the performance boost you need
Source: nvidia
DLSS 4 has a narrow window where it works best, and that comes up to its operation. Like the generation of framework in DLSS 3, MFG in DLSS 4 uses the interpolation of the frame. Two images are rendered and compared, then the weft generation algorithm uses this comparison to estimate the frame (or the frames) which should take place between the two. The more there is any difference between these two images, the more the algorithm of weft generation must operate, and it is the heart of the problem with MFG.
When you play a game at a low frequency of images, there is a lot of movement between two images. The weft generation algorithm works as best when there is only a small difference between two images, because it does not need to dream of details that may not be present in the two frames rendered. This is where MFG shows the most problems. You can see in Cyberpunk 2077 Below an example of what I’m talking about. With the rapid movement of a fight, the MFG turns into a mess of visual artefacts because it tries to follow the low frequency of basic images.
It is not a revelation – we know the limits of the framework generation for a certain time – but it is an illustration of the way in which the MFG can collapse outside its narrow use. You would like a 4x increase in the most image frequency when you start from a low basic images frequency, which is the situation where MFG is at worst.
If you have DLSS 3, you don’t need DLSS 4
Doubling your image frequency is quite good
DLSS 4 is excellent with regard to the super resolution and the generation of single frame, but the application of MFG is extremely narrow. If you have a too low basic images frequency, it disintegrates in a mess of visual artefacts. And if you already have a high basic images frequency, you don’t need to quadruple your image frequency, unless you are on the dating technology bleeding. In the vast majority of situations, the MFG simply has no meaning.
The generation of Nvidia DLSS frame is fantastic, and I used it generously in recent and demanding games like Avou, Indiana Jones and the large circle, And Monster Hunter Wilds. I never had the impression that I needed to generate three or four images instead of two. The useful window for the generation of 2x frame is much higher than the window for the 4x frame generation, and in the vast majority of the games I play, this is where I find myself landing.
As a functionality, DLSS MFG is not intrinsically problematic. Although its applications are narrow, there is a legitimate argument to use the functionality when you run a game at 60 IPS on a 240 Hz screen. As an additional goodie on which you can occasionally rock, MFG is an excellent feature to have. It is simply not a reason to go out and buy a new graphics card alone, and taking into account what we have seen from the GPUs of the NVIDIA RTX 50 series, it is the The main blackwell cards bring to the table.
Unlike DLSS 3 and its single frame generation capacities, DLSS 4 and MFG are not a reason to upgrade. I barely touched the functionality since the upgrading of my GPU, and I do not see it changing if soon.