These Nvidia App tweaks transformed my games, and most RTX owners never use them
At a glance:
- DLSS Override Model Presets can unlock newer AI models in older games.
- Smooth Motion doubles frame rates on RTX 40/50 series without per-game support.
- FPS caps can actually lower latency and improve responsiveness.
The Nvidia App's hidden power
The Nvidia App has quietly become one of the most important pieces of software on an RTX gaming PC, and most people using it have no idea. It replaced both GeForce Experience and the old Nvidia Control Panel over the past couple of years, and while you can access a lot of good stuff on the latter app alone, the Nvidia App does have certain settings that aren't available in the legacy control panel. Buried inside the app are settings that can meaningfully improve image quality, reduce latency, smooth out frame pacing, and even squeeze more life out of older cards. Some of them are driver-level overrides that work across hundreds of games without any per-title configuration, but a lot of these are going to be useful to test and enable on a per-game basis.
DLSS Override - Model Presets
Models completely change how your games look
This is the single most impactful toggle in the entire app, and it's one that most people have never touched. Games that support DLSS ship with a specific version of the model baked in, and that version is often whatever was current when the game launched. That means a game from 2023 is probably still running on an older CNN-based DLSS model, even though Nvidia has shipped dramatically improved transformer-based models since then. Overriding the model preset lets you force the latest available model globally or on a per-game basis. Setting this to the recommended preset gives you DLSS 4.5's current transformer model across hundreds of supported titles, which generally means noticeably cleaner image quality with less ghosting and better detail in motion.
A small asterisk is worth noting for owners of older RTX cards: RTX 20 and 30 series lack native FP8 support, which means that the newer models carry a performance hit. If you're on a card from one of those generations, staying on Preset K (DLSS 4) is usually the better balance of performance and image quality. Owners of RTX 40 and 50 series cards can more easily set-and-forget.
DLSS Override - Super Resolution Mode
DLAA in games that don't expose it natively
The second DLSS override that makes sense to use sits right next to the first one. It's where you decide how aggressively DLSS upscales, and the presets of Quality, Balanced, Performance, Ultra Performance are familiar to anyone who has used DLSS before. This isn't interesting on its own, but it's where you can force DLAA, which is where this setting is truly useful. DLAA uses the DLSS model at native resolution for anti-aliasing only, which means no upscaling, just AI-powered image cleanup. A lot of games don't expose DLAA as an in-game option, even when they support DLSS, and this override allows you to enable it. If you have GPU headroom and care about image quality more than raw frame rate, forcing DLAA globally is genuinely transformative in older titles. For newer, more demanding games where you need the performance, sticking with Quality or Balanced mode is the more optimal choice.
Smooth Motion
Frame interpolation for titles that don't support it
Smooth Motion is Nvidia's driver-level frame interpolation feature for RTX 40 and 50 series cards. It essentially allows you to double your framerate with very minimal visual impact. For older titles, smaller releases, and anything else that doesn't have native Frame Generation support, it's a solid way to increase performance. It's not as clean as native DLSS Frame Generation because the driver doesn't have access to the same motion vectors and game data, but it works on practically everything. For games like Escape From Tarkov where framerates are at a premium, the results really speak for themselves. It had actually felt like I had doubled my framerate, and I couldn't notice any visual artifacting, at least not at 4K.
It does force Low Latency Mode to Ultra, presumably to decrease the added latency that frame interpolation induces, and that's probably where the big differences in overall latency come from. The perceived framerate decrease that Low Latency Mode introduces is practically nonexistent, so that's a trade-off I'm willing to make, even in competitive titles.
FPS Cap
A useful way to make your games more responsive
Capping FPS sounds counter-intuitive, but in a lot of situations, it actually delivers a way more responsive experience than an uncapped framerate would, in addition to some other key benefits. In the Nvidia App, you can cap FPS on a per-game basis, but you can also create a global cap.
When I tested an FPS cap of 240 for Counter-Strike 2, I noticed much lower PC latency and Present API latency. PC latency is self-explanatory, but Present API latency refers to how long the CPU spends waiting on the graphics API to send a frame off to the display pipeline. It's a CPU measurement, but it gives us a good picture of what the FPS cap does. I also tested it alongside the in-game FPS cap that Counter-Strike 2 offers, because there used to be a pretty significant difference between the two approaches, but I didn't notice a large difference in this case.
The Nvidia App is a gold mine for tweaking
None of these settings are one-size-fits-all, and I'd recommend you test them yourself before applying them wholesale. In regard to the DLSS models, Nvidia pushes updates every few months, and the other features, like Reflex are receiving sequels and small updates as time goes on. The Nvidia App offers completely different settings to the legacy Nvidia Control Panel, and if you haven't played with any of the new toggles and have any generation of RTX card, I promise it's worth your while.
FAQ
What is the most impactful Nvidia App setting most users miss?
Can RTX 20 and 30 series cards use the newest DLSS models?
How does Smooth Motion work and when should I use it?
More in the feed
Prepared by the editorial stack from public data and external sources.
Original article