How Should You Cap Your Frame Rate? In-Game vs Driver vs RTSS Limiters, VRR Pairing & Frame-Pacing Comparison Guide
重點摘要
A frame-rate cap controls pacing, not raw speed. This guide compares in-game, driver-level and RTSS limiters on latency and frame-time stability, explains why only one limiter should run at a time, and shows how to pair a cap with VRR so frame rates stay inside the variable-refresh range.
A Cap Controls Rhythm, Not Power
A frame-rate limiter sets a ceiling on how many new frames a game hands over each second. It does not make a graphics card faster, it does not add rendering power, and it does not repair a game that is struggling to render. What it does is impose a steady rhythm on frame delivery. That distinction matters, because many players judge smoothness by the frame-rate number alone, when the feeling of smoothness depends far more on how evenly spaced each frame is.
Think of it this way. A game running at an average of 120 frames per second can still feel uneven if some frames arrive after a short gap and others after a much longer gap. The average looks good, but the experience stutters. Conversely, a game capped at a lower, steady rate can feel smoother than an uncapped game with a higher average, because every frame arrives at a predictable interval. This is why frame time — the gap between one frame and the next — is a better judge of a cap than the headline frame-rate number.
A cap also has practical side benefits. By stopping a game from rendering far more frames than a display can show, it can reduce power draw, heat and fan noise. In a game where the graphics card is easily producing more frames than needed, a sensible cap wastes less energy without changing what the player sees. But these are side effects of a steady ceiling, not the main purpose. The main purpose is pacing.
It is also important to separate the job of a limiter from the job of graphics settings. Graphics settings decide how much work each frame requires, and therefore how many frames a system can produce. A limiter decides the rhythm at which those frames are released. Lowering graphics settings may raise the ceiling a system can reach; a cap then decides whether to use all of that ceiling or to hold back for steadiness. The two tools work together, but they are not interchangeable.
The Three Kinds of Limiter
There are three common places to set a cap: inside the game itself, in the graphics driver, or in an external tool such as RivaTuner Statistics Server (RTSS). All three aim at the same result, but they sit at different points in the rendering pipeline, and that position creates different trade-offs between responsiveness and frame-time stability.
In-Game Limiters
An in-game limiter is controlled by the game engine itself. Because it sits closest to the rendering pipeline, it generally introduces the least additional delay of the three options. The engine knows when a frame is ready and can pace delivery without an extra layer of interception. For most players, this is the recommended first choice: if a game offers a frame-rate limit option, try it first.
The weakness of in-game limiters is consistency. Quality varies from game to game. Some engines implement a very accurate, stable cap; others implement a rough cap that still allows noticeable frame-time variation. A few older games offer a cap that does not behave reliably at all, or offer no cap option in the first place. An in-game limiter is therefore the best starting point, not a guarantee. If its frame-time curve looks uneven in practice, moving to another limiter is a reasonable next step.
Driver-Level Limiters
Driver-level limiters, such as the maximum frame rate setting in a graphics driver control panel, apply across games. Their strength is universality: one setting covers every title, including games that offer no built-in cap. This makes them a convenient fallback and a good choice for players who want one consistent rule rather than per-game configuration.
In general, a driver-level limiter tends to add slightly more delay than a well-implemented in-game limiter, because it sits one step further from the engine. The difference is usually described as small rather than dramatic, and for non-competitive play it is rarely the deciding factor. The benefit is management: a single place to set, change and remove the cap.
External Limiters: RTSS
RTSS is an external frame-rate limiter that has built a reputation for exceptionally steady frame-time curves. Players who prioritise the smoothest possible frame delivery often choose it for exactly that reason: once set, the spacing between frames tends to stay very even.
The trade-off is that an external limiter intercepts frames from outside the engine, adding an extra layer to the pipeline. As a result, its delay tends to be the highest of the three options in general terms. For single-player, story-driven or visually focused games, that trade-off is often worth it. For highly competitive play, where every increment of responsiveness matters, it deserves more careful thought. RTSS is a tool with a clear strength and a clear cost, not a universal best answer.
| Limiter | Where it sits | General strength | General trade-off | Best suited to |
|---|---|---|---|---|
| In-game limiter | Inside the game engine | Lowest added delay in general | Accuracy varies by game | Games that offer a reliable built-in cap |
| Driver-level limiter | Graphics driver | Works across all games, one place to manage | Slightly more delay than in-game in general | Games with no built-in cap, or players who want one rule |
| RTSS external limiter | Outside the engine | Very steady frame-time curve | Highest added delay of the three in general | Players who prioritise smoothness over minimum delay |
One Limiter at a Time: The Iron Rule
Only one limiter should be active at any time. Running two caps at once — for example, an in-game cap together with a driver cap, or a driver cap together with RTSS — does not make pacing more stable. Each limiter tries to impose its own rhythm, and the two rhythms interfere with each other. The result is typically a less stable frame-time curve than either limiter would produce alone.
This is one of the most common setup mistakes, because it feels intuitive to add more control when something feels uneven. In practice, the fix for an uneven curve is to pick one limiter, test it properly, and only switch to a different limiter if the first one proves unsuitable — never to stack them. If a cap is set in a game, the driver cap should be off. If RTSS is in use, both the in-game and driver caps should be off. Keep the rule simple and it will prevent a whole class of confusing stutter.
Choosing the Cap Value: There Is No Single Magic Number
The right cap value depends on the display, on whether variable refresh rate (VRR) is in use, and on what the game can sustain. A single number recommended for every situation would be misleading. The sections below give a scenario-based approach instead.
With VRR (G-Sync or FreeSync)
With VRR active, the goal is to keep the frame rate inside the display's variable-refresh range at all times. If the frame rate rises above the display's maximum refresh rate, it leaves that range, and the system falls back to behaviour closer to traditional synchronisation, with its associated delay. A cap exists, in this scenario, to prevent that overshoot.
Public guidance from display-testing sources, including Blur Busters G-SYNC 101, suggests setting the cap a few frames below the display's maximum refresh rate — described as a range of about two to three frames below the maximum — to leave room for differences in limiter accuracy. For a 144 Hz display, that means a cap in the region of the low 140s rather than exactly 144. The point of the small margin is not precision for its own sake: different limiters are accurate to different degrees, and a small buffer keeps the frame rate safely inside the VRR range even when a limiter overshoots slightly. Setting the cap exactly at the maximum refresh rate risks repeated excursions above it; setting it far below wastes refresh headroom for no benefit.
With VSync and No VRR
Without VRR, and with vertical synchronisation (VSync) enabled, the interaction is different. VSync holds frames to the display's refresh cycle, and a cap set exactly at the refresh rate can interact with that waiting behaviour in ways that feel uneven. The detailed behaviour of VSync, tearing and stutter belongs to a separate discussion of display synchronisation, but the practical point here is simple: a cap and VSync are two pacing mechanisms that must be understood as a pair, not set independently and forgotten.
Without VRR and Without VSync
Without VRR and without VSync, there is usually no smoothness reason to cap at all, unless the goal is to reduce power draw, heat or fan noise. In competitive play, an uncapped frame rate is often preferred, because a higher frame rate generally means fresher information reaching the display and lower input delay. Capping in this scenario trades responsiveness for quietness or coolness — a legitimate choice, but one that should be made knowingly.
When Performance Is Limited
A cap should sit near a level the game can sustain over time. If a game typically runs at around 50 to 60 frames per second in demanding scenes, setting a cap of 120 does nothing: the cap is never reached, so it imposes no rhythm at all. Conversely, setting a cap far below what the game sustains creates a different problem, covered in the myths section below. The useful cap is one the system actually hits consistently, turning a variable frame rate into a steady one.
How to Tell Whether a Cap Is Working
The test is straightforward and does not require expensive tools. Monitoring software that can plot a frame-time graph is enough, and such tools are generally available free of charge.
Run the same scene twice: once with no cap, once with the chosen cap. Compare the frame-time curves, not just the average frame-rate numbers. Before the cap, the curve may show large spikes and uneven spacing. After a well-chosen cap, the curve should look flatter and more regular. If the frame-rate number barely moves but the game feels smoother, that is a normal and expected result — the cap was never meant to raise the number, only to steady the rhythm.
If the curve does not improve, check the iron rule first: is another limiter still active somewhere? Then check whether the cap value is actually being reached. A cap set above the game's sustained performance cannot steady anything. Only after those two checks should a different limiter be tried.
How a Cap Interacts with Other Settings
A limiter is one part of a display pipeline, and its behaviour depends on its neighbours. With VRR, the cap, the VRR range and the VSync setting form a group that should be considered together; changing one without thinking about the others can undo the benefit of the cap. Low-latency modes that manage the render queue, such as NVIDIA Reflex or AMD Anti-Lag, are a separate layer again: they control how deep the queue of prepared frames grows, while a limiter controls the rhythm at which frames are released. The two can coexist and address different parts of delay, but they should not be confused with each other.
Frame generation adds another boundary. When frame generation is enabled, a cap applies to the final displayed frame rate, after generated frames are included, not to the underlying rendered frame rate. That changes how a cap value should be interpreted, and it means a cap chosen for a non-generated game may not translate directly to a generated one.
Finally, a cap has clear limits. It will not fix stutter caused by shader compilation, by running out of video memory, or by a processor bottleneck. Those are different problems with different causes, and a cap only controls pacing. If stutter persists with a steady frame-time curve, the cause lies elsewhere, and chasing the cap value further will not help.
Common Myths, Dismantled
Myth: Higher FPS is always smoother. A high average with uneven frame spacing still feels uneven. Judge the frame-time curve, not the headline number.
Myth: Two caps are more stable than one. Stacked limiters interfere with each other and make frame timing less stable, not more. Use exactly one.
Myth: A cap makes a computer faster. A cap cannot raise performance. It steadies rhythm and may reduce power draw and heat, but it adds no speed.
Myth: A VRR display needs no cap. Without a cap, frame rates can overshoot the top of the VRR range and fall back to synchronisation-style delay. A cap just below the maximum refresh rate is what keeps the frame rate inside the range.
Myth: One limiter is always best. There is no absolute winner. The lowest-delay option in general is the in-game limiter; the steadiest curve in general comes from an external limiter; the most convenient across games is the driver limiter. The right choice depends on whether the priority is responsiveness, smoothness or convenience.
Myth: The lower the cap, the smoother the game. An excessively low cap directly increases the gap between frames and increases input delay, which is especially noticeable on lower-refresh-rate displays. Smoothness comes from steadiness at a sensible level, not from the lowest possible number.
Edge Cases Worth Knowing
Some older games have built-in caps that are inaccurate or ineffective. In that case, switching to a driver or external limiter is a reasonable fallback rather than persisting with a broken in-game option. When frame generation is active, remember that the cap counts displayed frames, including generated ones. On a multi-monitor setup with different refresh rates, the cap should be based on the refresh rate of the display the game is actually running on. On a laptop, a cap can also serve as a cooling and battery-saving measure, although laptop power behaviour is a separate topic with its own constraints.
Frequently Asked Questions
Should I always use the in-game limiter? Try it first. If the game offers one and its frame-time curve looks steady, there is usually no reason to switch. Move to a driver or external limiter only if the in-game option is missing, inaccurate or unstable.
Can I use RTSS and a driver cap together for extra smoothness? No. Only one limiter should be active at a time. Stacking limiters makes frame timing less stable.
What cap should I use on a VRR display? As a starting point, a few frames below the display's maximum refresh rate — roughly two to three frames below — keeps the frame rate inside the VRR range while allowing for limiter accuracy differences. Adjust from there based on the observed frame-time curve.
Why does my game feel smoother even though the FPS number did not change? Because the cap changed the spacing between frames, not the average count. Steadier spacing feels smoother even at the same average frame rate.
Will a cap fix stutter in every game? No. A cap only controls pacing. Stutter from shader compilation, video memory pressure or a processor bottleneck needs a different diagnosis and a different fix.
Does a cap help on a laptop? It can reduce heat, fan noise and power draw, and it can steady pacing just as on a desktop. Laptop power limits are a separate factor and should be considered alongside the cap, not confused with it.
When to Seek Professional Help
If frame pacing remains uneven after trying one limiter at a time, checking the cap value against sustained performance, and confirming that no second limiter is active, the cause may be hardware-related — for example, thermal throttling, a failing component or a deeper system issue. In that case, a professional inspection can separate a settings problem from a hardware problem before parts are replaced unnecessarily.
HK GPU Fix is located at Dali Plaza, Room 1209, Shek Mun. For display or graphics-related inspection, call 2186 8010 or reach us on WhatsApp at 6558 6806. Inspection and repair fees follow the published three-tier structure of HK$400 to HK$2,800, depending on the work required. Note that we do not operate a Mong Kok shop, and we do not offer a door-to-door service; visits are by appointment at the Shek Mun workshop.