VSync, G-SYNC and FreeSync

By Walter Jobs, Technology writer and editor · Published 10 October 2026

One piece of advice dominates this subject and it is wrong as stated: turn VSync off if you have G-SYNC. NVIDIA’s own latency guide documents the opposite configuration as supported, names it, and explains what it does. We will start there, because if you take nothing else from this page, take the correction.

Then three things nobody else covers: that the three organisations certifying variable refresh displays each specify the property differently, what actually happens when your frame rate falls below the panel’s window, and where a widely repeated number about that has no source behind it.

The correction, in NVIDIA’s words

The advice did not come from nowhere. NVIDIA does say this, in general terms: “Turn off VSYNC – The age old latency optimization; turn off VSYNC. VSYNC causes back pressure from the display that reverberates through the entire system. In general, we highly recommend turning VSYNC OFF if you are willing to tolerate tearing.”

That last clause is the one that gets dropped. And on the same page, NVIDIA states the variable refresh exception explicitly: “However, if you have a variable refresh rate display, like an NVIDIA G-SYNC monitor, you can get the best of both worlds: no tearing (if your FPS is below your refresh rate), and no VSYNC latency.”

Then the sentence that settles it, naming the configuration and what it does:

“For G-SYNC gamers who don’t want to tear, keeping VSYNC ON while using NVIDIA Reflex or NVIDIA Ultra Low Latency Mode, will automatically cap the framerate below the refresh rate, preventing VSYNC backpressure, eliminating tearing, and keeping latency low if you become GPU bound below the refresh rate of your display. Do note, however, that this method will result in slightly higher latency than just letting your FPS run uncapped with NVIDIA Reflex enabled.”

So VSync on, with a latency feature, on a variable refresh display, is not a mistake. It is a configuration NVIDIA documents, and the mechanism is that it imposes a frame cap for you rather than waiting on the display. NVIDIA presents it as a trade: tearing gone, latency slightly higher than uncapped. That is a choice, not an error, and anyone telling you flatly to turn VSync off with G-SYNC is quoting half of NVIDIA’s sentence.

One footnote NVIDIA adds that matters in practice: “VSYNC ON in the NVIDIA Control Panel will only work for Fullscreen applications. In addition, MS Hybrid-based laptops do not support VSYNC ON. If you are gaming in windowed mode or on one of these laptops, and want to utilize G-SYNC + VSYNC + Reflex mode, use in-game VSYNC.”

What each technology actually is

Worth separating, because the three words in the title are not three brands of the same thing.

VSync is the old solution and it works by making the game wait. NVIDIA’s definition: it “allows you to control Vertical Synchronization mode, where the application’s frame rate is synchronized with the display refresh rate in order to eliminate tearing”. The display runs at a fixed rate and the game is held back to match it. That removes tearing and introduces waiting, which is the back pressure in the quote above.

Variable refresh inverts the relationship: instead of the game waiting for the display, the display waits for the game. AMD’s description of how the range is established: “FreeSync capable graphics cards can detect and set the appropriate maximum and minimum refresh rate based on the capabilities reported by the display. This means FreeSync supports up to the maximum refresh rate of the display itself.”

The difference between the branded implementations is mostly about whether there is dedicated hardware in the monitor. NVIDIA’s tiers make this explicit. Its G-SYNC tier “features a NVIDIA G-SYNC processor” and promises “a full variable refresh rate (VRR) range and variable overdrive”. Its third tier is the one most people own: “G-SYNC Compatible – Doesn’t use NVIDIA processors but is validated by NVIDIA to deliver a good, basic variable refresh rate (VRR) experience for tear-free, stutter-free gaming.”

AMD took the opposite position from the start, and says so: “FreeSync technology keeps costs low by not requiring an expensive module in the monitor to replace the standard monitor scaler.” On the overdrive question, where NVIDIA’s module tier makes a hardware claim, AMD’s answer is that its standard permits it without demanding it: “AMD FreeSync technology has always been fully compatible with variable overdrive”, and “as such we do not require, but do encourage, monitor OEMs to implement variable overdrive”.

Three organisations, three different ways of specifying the same thing

This is the part we have not seen covered anywhere, and it explains why monitor specifications in this area are so hard to compare. Each certifying body pins down the variable refresh range differently.

A note on VESA’s figures, because they get misused: those are logo requirements for a certification programme, not statements about human vision. VESA publishes no perceptual threshold for frame rate anywhere. Its certification does test things most people assume nobody checks, including flicker and overdrive overshoot: its specification covers “refresh rate, flicker, gray-to-gray response time (including limits on overshoot and undershoot to ensure high-quality images), video frame drop, and video frame rate jitter”, across “more than 50 automated display performance tests”, with every display “tested in the factory shipping state or default factory mode configuration”. There is now a third tier as well, AdaptiveSync Dual Mode, for displays offering a high-resolution and a high-refresh mode.

NVIDIA’s validation criteria are worth quoting in full too, because the ratio is only part of it: “G-SYNC Compatible testing validates that the monitor does not show blanking, pulsing, flickering, ghosting or other artifacts during VRR gaming. They also validate that the monitor can operate in VRR at any game frame rate by supporting a VRR range of at least 2.4:1 (e.g. 60Hz-144Hz), and offer the gamer a seamless experience by enabling VRR by default.”

AMD states that certification happens on its side too: “Every AMD FreeSync display goes through a rigorous certification process, testing its refresh rates, latency, HDR color and luminance accuracy and much more.”

Below the window: the thing that actually goes wrong

Every variable refresh display has a bottom to its range. Drop below it and the display cannot slow down any further, and what happens next is where the two vendors diverge sharply, in documentation if not necessarily in hardware.

AMD documents a mechanism by name, and the full explanation is the clearest thing either vendor has published on this subject:

“Low framerate compensation (LFC), allows AMD FreeSync technology to work when the framerate falls below the minimum refresh rate of the display. When the framerate drops below the minimum refresh rate of the display, frames are duplicated and displayed multiple times so that they can sync to a refresh rate that is within the displays refresh rate range. For example, a display with a 60 – 144Hz refresh rate, would be able to sync the frames of a game running at 40 FPS, by doubling them so that the display could sync and run at 80 Hz. A display with LFC effectively results in the removal of the minimum refresh rate boundary. All displays in the AMD FreeSync Premium and FreeSync Premium Pro tier are certified to meet mandatory LFC requirements.”

Frames shown twice to keep the panel inside its range. The 40 fps game drives an 80 Hz panel, which is inside the window, so the sync keeps working. That is the whole trick, and it is why the frame rate floor of a display matters less than its specification sheet implies.

Two honest complications.

AMD’s own pages do not agree. Its tier comparison table marks low framerate compensation with a tick in all three tiers, including the baseline. The FAQ on the same page restricts the certified mandate to two: “All displays in the AMD FreeSync Premium and FreeSync Premium Pro tier are certified to meet mandatory LFC requirements.” We cannot reconcile that for you. If this feature is the reason you are buying a particular monitor, the Premium tiers are the ones where AMD commits to it in writing.

NVIDIA documents no equivalent. We looked for a statement of what happens below the range on its hardware and did not find one. What NVIDIA does state is that its module-equipped displays “feature a full refresh rate range from 1Hz to the display panel’s max refresh rate”, which is a different claim: not a frame-doubling mechanism but a range with effectively no bottom. For its validated third-party panels, the 2.4:1 ratio requirement is NVIDIA’s answer to the same problem, expressed as a specification rather than a technique. We are not going to describe a mechanism NVIDIA has not published.

A number to stop repeating

You will frequently read that low framerate compensation requires a display whose maximum refresh is at least twice, or two and a half times, its minimum. We went looking for the source. AMD publishes no such multiplier and no ratio condition anywhere on its FreeSync pages. The only numeric thing it gives is the worked example above, a 60 to 144 Hz panel doubling 40 fps to 80 Hz, which is an illustration rather than a requirement.

Our reading of where the figure came from, offered as a reading: NVIDIA’s separate 2.4:1 validation rule for G-SYNC Compatible panels is a real, published ratio about variable refresh ranges, and it has very likely been absorbed into writing about AMD’s feature over the years. It is a sensible rule of thumb either way, since a range narrower than 2:1 leaves no room to double a frame rate into. But attributing it to AMD is attributing something AMD has not said.

One more piece of stale information to watch for: an older AMD support article still publishes the 2020 tier criteria, including a 120 Hz figure and a minimum resolution requirement that the current FreeSync page does not contain. If you find two different sets of AMD tier requirements, the product page is the current one.

What to set

Three configurations cover almost everybody. Pick by what you care about.

If you want no tearing and you have a variable refresh display. Turn on the variable refresh in the driver, turn on the game’s latency feature if it has one, leave VSync on, and set a frame cap a few frames below your panel’s maximum. This is the configuration NVIDIA names, and the cap keeps you inside the window so the VSync never has to engage. Blur Busters, from its own testing, recommends “a minimum of 3 FPS limit below display’s maximum refresh rate”, with 141 at 144 Hz as its worked example, and notes that in this configuration VSync “will never engage”. That is their tested recommendation rather than a vendor specification, and we cite it as such.

If you want the lowest possible latency and will accept tearing. VSync off, latency feature on, no cap. NVIDIA’s guide says plainly that this is lower latency than the configuration above, and its general line applies: turn VSync off “if you are willing to tolerate tearing”. Add exclusive fullscreen, which NVIDIA also recommends: “If possible, always be in Exclusive Fullscreen mode. This will bypass the windows compositor that adds latency.”

If you have a fixed refresh display. You are choosing between tearing and waiting, and there is no third option. VSync on removes the tearing and adds latency. VSync off does the reverse. NVIDIA’s adaptive modes, which engage VSync only above the refresh rate, are a middle path and are available only on a fixed refresh display: NVIDIA documents that its Adaptive setting “is available only if Monitor Technology is set to ULMB or Fixed Refresh Rate”. Which is to say it is not an option on a G-SYNC configuration, which is the correct design, because variable refresh is already solving that problem.

Whichever you choose, whether to cap and where is a decision with its own complications, including a place where NVIDIA’s documentation disagrees with itself, and we cover that separately in should you cap your fps. For why the latency features matter more than any of these settings, see frame rate and input lag. And to know whether your hardware will keep you inside your panel’s window in the first place, the fps calculator is the place to start.

Common questions

Is G-SYNC better than FreeSync?

A monitor with NVIDIA’s hardware module gets a wider guaranteed range and variable overdrive from the vendor’s own claim. Beyond that, the two standards do the same job, and the quality differences between individual monitors are larger than the differences between the badges. Both vendors run certification programmes; check which badge the specific panel carries rather than comparing brands.

Can I use FreeSync with an NVIDIA card?

That is what the G-SYNC Compatible tier is: panels not using NVIDIA’s module that NVIDIA has validated. NVIDIA also lets you enable it on unvalidated displays, with a candid warning: “it may work, it may work partly, or it may not work at all.”

Should VSync be on in the game or in the driver?

If you are using the configuration NVIDIA names, the driver setting is the one its guide describes, with the caveat that it works only in fullscreen and not on hybrid-graphics laptops, where NVIDIA says to use the in-game setting instead.

Why does my variable refresh display still stutter?

Two likely reasons. Your frame rate may be falling out of the bottom of the panel’s window, which is what the section above is about. Or the stutter is frame-time variance rather than a sync problem, which variable refresh cannot fix because it is not a sync fault. 1% lows is how you tell the two apart.

Do I still need VSync at all in 2026?

On a variable refresh display, as a safety net for the moments your frame rate overshoots the window, which is exactly the role NVIDIA’s guide describes it in. On a fixed refresh display it remains the only way to eliminate tearing.