GPU Screen Tearing vs Stutter: VSync, G-Sync/FreeSync, Frame Caps & Frame Pacing Explained
Key Takeaways
Tearing puts two frames in one refresh, while stutter is uneven frame delivery. This guide shows how to tell them apart from symptoms and frame-time graphs, then set VSync, VRR and frame caps correctly for your display without guessing or changing several settings at once.
Introduction: Why "Not Smooth" Can Mean Several Different Things
When a game or video does not feel smooth, it is tempting to describe everything with one word. In practice, several distinct problems hide behind that description. A horizontal line where the upper and lower halves of the picture do not line up is tearing. A picture with no split line, but movement that hesitates and jumps, is usually stutter or uneven frame pacing. A moving object that leaves a faint trail behind it is more likely a display response-time or overdrive issue. A mouse that feels half a step behind your hand is an input-latency issue.
Treating all of these as the same problem leads to random troubleshooting: turning on VSync removes tearing but makes stutter worse, buying a variable refresh rate display brings no benefit because the settings conflict, or a high average frame-rate number still feels rough to the eye. This article separates the phenomena first, then explains what VSync, G-Sync and FreeSync, and frame caps each control, and finally gives a step-by-step diagnostic process you can follow yourself.
One note before starting: this is an educational and diagnostic guide about general mechanisms and settings logic. Menu names, driver behaviour, and game options vary between displays, driver versions, and titles, so always check the documentation for the hardware and software in front of you.
Part 1: Telling Them Apart — Tearing, Stutter, Ghosting, and Input Lag
What tearing looks like
Tearing is most visible when the camera pans quickly sideways. A horizontal boundary appears somewhere in the picture, and the scenery above and below that line is offset sideways. When nothing moves, tearing is usually invisible, because consecutive frames differ very little. When the camera moves fast, consecutive frames differ a lot, and the offset becomes obvious.
The defining feature of tearing is that more than one frame appears within a single refresh cycle. It is not slowness and it is not a pause. It is a picture cut into pieces from different moments. It is most likely when all synchronisation is off and the graphics card outputs frames without regard to the display refresh rhythm.
What stutter looks like
Stutter has no split line. Objects remain whole, but their movement is uneven. Walking feels as if something briefly holds you back and then lets go. Panning the camera feels like a series of small jumps rather than a continuous turn. Some stutter is intermittent: smooth for a few seconds, then a sudden hitch, then smooth again. Some is a persistent micro-stutter that makes the whole session feel slightly rough.
The defining feature of stutter is uneven delivery timing. Even if enough frames are produced within a second, the gaps between them matter. If those gaps swing between short and long, the eye notices the broken rhythm. Frame time, discussed in detail later, is the most useful tool for identifying this kind of problem.
Ghosting and input lag are different again
Ghosting is a faint blurred trail behind a fast-moving object. It is usually related to panel response time and the overdrive setting in the display menu. Overdrive set too aggressively can produce inverse ghosting, a bright or dark halo around moving edges. Overdrive set too weakly leaves a longer smear. These are display-menu issues and should not be confused with graphics-card synchronisation settings.
Input lag is the time between pressing a key or moving the mouse and seeing the result on screen. Synchronisation settings, frame caps, and buffering depth all affect it. Competitive players are especially sensitive to it, but many general users describe high input lag as stutter, when the picture itself may be moving evenly. During diagnosis, keep "the picture is uneven" and "the response feels late" as two separate questions.
| Symptom | What you see | Common trigger | First area to suspect |
|---|---|---|---|
| Screen tearing | A horizontal split line; upper and lower halves misaligned | Fast camera movement, frame rate far above refresh rate, synchronisation off | Synchronisation settings; mismatch between GPU and display rhythm |
| Stutter / micro-stutter | Whole picture, but movement hesitates and jumps | Sudden scene complexity, background programs competing for resources, uneven frame times | Frame pacing, system load, setting conflicts, storage read pauses |
| Ghosting / inverse ghosting | Blurred trail or bright halo behind moving objects | Fast motion, unsuitable overdrive setting | Display response time and overdrive strength |
| High input lag | Actions feel half a step behind your hand | Synchronisation on, multiple buffer layers, very low frame cap | VSync, buffering, triple buffering, frame-cap settings |
Part 2: How a Display Draws One Picture
To understand tearing, it helps to picture how a display works. A display does not print the whole picture at once. It scans from top to bottom, row by row, sending the contents of the frame buffer to the panel. One complete scan is one refresh. The refresh rate is how many scans happen per second: 60 Hz means sixty scans per second, and 144 Hz means one hundred and forty-four scans per second.
On the graphics-card side, there is a division of labour between a front buffer and a back buffer. The front buffer holds the picture the display is currently reading. The back buffer is where the graphics card draws the next picture. When the back buffer is finished, the two buffers swap during the vertical blanking interval, the short gap between the end of one scan and the start of the next. Swapping in that gap means the next scan reads one complete picture from start to finish, with no mixture of old and new frames.
Tearing happens when that swap timing goes wrong. If the graphics card replaces the contents of the front buffer in the middle of a scan, the upper part of the display shows the old picture and the lower part already shows the new picture, and the two halves do not line up at the boundary. The faster the graphics card outputs frames relative to the refresh rhythm, the more likely it is that a single scan will contain pieces of more than one frame, and there may be more than one tear line.
Every synchronisation method is, at heart, an answer to one question: who should adapt to whose rhythm? Traditional VSync answers that the graphics card should adapt to the display. Variable refresh rate answers that the display should adapt to the graphics card. A frame cap answers that the graphics card should first be kept within a predictable range, so the two rhythms conflict less often. Each answer has a cost, and the following sections examine them one by one.
Part 3: VSync — The Oldest Method, and the Most Misunderstood
How VSync works
With VSync enabled, the graphics card only swaps buffers during the vertical blanking interval, so every scan reads a complete picture. The cost is waiting: if a frame is not ready by the next blanking interval, the display has no choice but to repeat the previous frame until the new frame is finished. That waiting time is a major source of VSync input lag.
On a fixed-refresh display, the effect is straightforward. Tearing disappears, but the frame rate becomes locked to the refresh rate or to a whole-number fraction of it. On a 60 Hz display, for example, if the graphics card occasionally cannot finish a frame in time for the next refresh, repeated frames appear and smoothness changes in steps rather than gradually. This stepped behaviour is especially noticeable with older double-buffered implementations.
Double buffering versus triple buffering
Double buffering uses one back buffer. After drawing a frame, the graphics card must wait for the swap to complete before it can start drawing the next frame. If the swap has to wait for the next blanking interval, the graphics card sits idle during that wait. Once the frame rate falls below the refresh rate, idle time grows and repeated frames become more frequent.
Triple buffering adds a second back buffer. While waiting for a swap, the graphics card can already start drawing the frame after next, reducing idle time. In practice, the frame rate falls less abruptly when it drops below the refresh rate, and the rhythm feels somewhat smoother. Triple buffering is not free, however: it uses more video memory, and because the graphics card may have drawn ahead based on older input, input lag is not necessarily lower than with double buffering. Implementations also differ between games and drivers, so enabling triple buffering in one place does not guarantee the same behaviour as enabling it elsewhere.
When VSync helps, and when it does not
If your display does not support variable refresh rate, and tearing bothers you more than input lag, VSync is the simplest direct solution. For video playback and slower-paced single-player games, the stable, tear-free picture is usually worth a small amount of added delay.
If you play fast-reaction competitive games, or if your display already supports variable refresh rate, traditional VSync alone may not be the best choice. Used on its own without a frame cap, VSync can produce repeated-frame judder when the frame rate fluctuates. In a variable refresh rate setup, VSync should step back and handle only the out-of-range edge cases, as explained in the combination section below.
| Approach | Method | What it solves | Main cost | Best suited to |
|---|---|---|---|---|
| No synchronisation | GPU outputs at full speed, ignoring refresh rhythm | Lowest input lag | Most visible tearing | Competitive play where tearing is acceptable |
| Double-buffered VSync | Swap buffers only during blanking interval | Eliminates tearing | Stepped frame-rate drops, higher lag | Fixed-refresh display, picture integrity over reaction speed |
| Triple-buffered VSync | Extra back buffer reduces GPU idle time | Eliminates tearing, softer frame-rate drops | More video memory, lag not necessarily lower | Fixed-refresh display with frame rate fluctuating around refresh rate |
| Adaptive synchronisation | VSync on below refresh rate, off above it | Balances tearing and frame-rate drops | Behaviour switches with frame rate, rhythm may feel inconsistent | Transitional solution, not ideal for users seeking a steady rhythm |
Part 4: G-Sync and FreeSync — Making the Display Follow the GPU
The variable refresh rate idea
Variable refresh rate, abbreviated as VRR, reverses the traditional logic. Instead of scanning at a fixed rhythm, the display waits for the graphics card to finish a frame and then refreshes. The graphics card produces a frame, the display refreshes once, and the two rhythms stay aligned moment by moment. This solves two problems at once: the picture is never replaced mid-scan, so there is no tearing, and the display is never forced to repeat an old frame while waiting, so there is no stepped frame-rate drop of the VSync kind.
G-Sync is NVIDIA's variable refresh rate technology. FreeSync is AMD's technology, built on the VESA Adaptive-Sync standard. The principles are similar, but certification and implementation differ. Many modern displays support both ecosystems, so the brand label alone is not a reliable guide. The actual specification and certification level matter more.
The VRR range: the number most often overlooked
Every VRR display has an effective range, usually written in the specification sheet as an interval such as 48 to 144 Hz. The meaning is simple: only when the graphics card's output rate falls inside this interval can the display follow it with a variable rhythm. Above the top of the range, the display cannot refresh any faster, and synchronisation falls back to traditional behaviour. Below the bottom of the range, the display cannot slow down any further, and synchronisation also stops working in VRR mode, unless a low-framerate compensation mechanism steps in.
This explains why two displays that both claim FreeSync support can feel very different. A model with a lower floor covers a wider range of frame rates. A model with a narrow range loses synchronisation earlier when the frame rate falls. When choosing a display, the range numbers deserve more attention than the badge on the box.
Low framerate compensation: rescuing low frame-rate scenes
Low framerate compensation, abbreviated as LFC, handles the case where the frame rate falls below the bottom of the VRR range. It repeats each frame so that the rate sent to the display returns to inside the range. If a display range starts at 48 Hz and the game output falls to a lower level, LFC can show each frame twice, bringing the display refresh rhythm back inside the range and keeping the picture in sync instead of suddenly reverting to fixed-refresh behaviour.
LFC is not guaranteed on every FreeSync display. FreeSync Premium requires LFC, while base FreeSync does not guarantee it, so the individual model specification must be checked. Native G-Sync module displays use a dedicated module inside the display, which allows a very wide range, with low frame-rate handling managed by the module itself. G-Sync Compatible refers to certified Adaptive-Sync displays, whose actual range and LFC behaviour depend on the individual model. In all cases, the conclusion is the same: read the range and LFC notes in the specification sheet rather than judging by the logo alone.
Compatibility checklist
Variable refresh rate only works when the whole chain cooperates. A break at any link reduces the benefit. Common checkpoints include: the display itself supports VRR and the relevant option is enabled in its on-screen display menu; the cable and connector support VRR, commonly DisplayPort, with some models also supporting HDMI VRR, subject to that model's documentation; the driver option for the corresponding technology is enabled and not overridden by a game setting; the game runs in fullscreen mode or in a mode that supports VRR; and the frame rate stays inside the display VRR range.
Multi-display setups deserve special attention. If the main display supports VRR but a secondary display runs at a fixed refresh rate, behaviour when switching between fullscreen and windowed modes may differ from a single-display setup. When this is suspected, testing with only the main display connected, as a comparison, can quickly show whether the multi-display combination is involved.
| Technology | Basis | Range characteristics | Low frame-rate handling | What to check before buying |
|---|---|---|---|---|
| G-Sync (native module) | Dedicated NVIDIA module | Usually very wide | Managed inside the module | Module version, range, additional features such as HDR behaviour |
| G-Sync Compatible | VESA Adaptive-Sync | Varies by model, wide or narrow | Depends on whether the model provides LFC | Certification listing, actual range, LFC notes |
| FreeSync | VESA Adaptive-Sync | Varies by model | Base level does not guarantee LFC | Actual range, LFC notes |
| FreeSync Premium | VESA Adaptive-Sync | Higher refresh-rate threshold required | LFC required | Lower end of range, whether LFC is stated |
| FreeSync Premium Pro | VESA Adaptive-Sync | Similar to Premium, plus HDR requirements | LFC required | Range, VRR behaviour in HDR mode |
Part 5: Frame Caps and Frame Pacing — Why Average FPS Can Mislead
Average FPS versus frame time
Average FPS counts how many frames were produced in one second. It is a total. Frame time measures how long each individual frame lasted before the next one arrived. It is a rhythm. The eye is far more sensitive to rhythm than to totals.
The conversion is simple: 60 FPS ideally means about 16.67 milliseconds per frame, and 144 FPS ideally means about 6.94 milliseconds per frame. The word "ideally" matters, because no real game produces every frame in exactly the same time. If most frames stay close to the ideal value, motion feels smooth. If occasional frames deviate sharply from the ideal, the eye notices a hitch, even when the average number still looks good.
Consider two sessions that both report an average of 60 FPS. In one, every frame lasts close to 16.67 milliseconds. In the other, most frames are fast and a few are very slow. The averages may look similar, but the second session will clearly feel worse. That is why detailed performance testing looks at frame-time graphs and low-percentile values in addition to averages: low-percentile values describe the worse stretches of performance, not the overall mean.
What a frame cap actually does
A frame cap does not make a computer run faster. It makes the rhythm more predictable. In a variable refresh rate setup, setting the cap slightly below the refresh rate, commonly a few frames below it, helps avoid repeatedly crossing the top of the VRR range. When the frame rate bounces across the range boundary, the synchronisation mode keeps switching, and the rhythm becomes less stable. A cap slightly below the ceiling draws a safety line under the boundary, letting synchronisation stay steadily inside VRR mode.
A second benefit is avoiding wasted work. When the frame rate runs far above what the display can show, the extra frames are never fully displayed, yet they still consume processing resources and generate heat. A sensible cap lets the graphics card work at a steadier pace, which can indirectly help frame-time stability. It is important to stress, however, that a cap alone cannot repair unstable frame times. If the fluctuation comes from sudden scene load, shader compilation on first encounter, or background programs, capping only pins the unstable rhythm at a lower level. The source of the fluctuation still needs to be found.
In-game caps, driver caps, and third-party caps: which one wins
Many users unknowingly run several cap layers at once: one in the game menu, one global limit in the driver, and one in a third-party tool. In practice, the strictest layer usually takes effect, but latency characteristics differ between layers. An in-game limiter generally adds less delay, because it can intervene earlier in the rendering process. A driver-level limiter is more universal but may add slightly more delay. A third-party tool depends on its implementation.
The recommended approach is to keep only one cap layer active and set the others to unlimited or off, then verify the effect with a frame-time graph. If changing one setting produces no visible change, another layer is probably taking effect first. It is not that the setting is useless; it is that a different layer is already in control. Changing one variable at a time is the only reliable way to identify which layer matters.
Where low-latency technologies fit
Low-latency technologies such as NVIDIA Reflex and AMD Anti-Lag reduce the queue between input and display, so the graphics card does not draw too many frames ahead. They can be used together with a frame cap in supported games, but they are not supported everywhere, and enabling them does not automatically fix stutter. If the root cause is uneven frame pacing, a low-latency technology can improve responsiveness without necessarily making the rhythm even. Responsiveness and rhythm should be evaluated separately.
Part 6: Setting Conflicts — The Most Common "Everything Is On, But It Still Feels Wrong"
In-game VSync plus driver VSync at the same time
This is one of the most common conflicts. Both the driver and the game have their own VSync switches. With both enabled, actual behaviour can become hard to predict: sometimes it amounts to double waiting, sometimes one layer overrides the other, and the result is extra delay or an unstable rhythm that is invisible from the menus alone.
A steadier approach is to let only one layer control synchronisation. A common VRR combination is: driver VSync on, in-game VSync off, plus a frame cap slightly below the refresh rate. In this combination, VRR leads inside the range, driver VSync handles the above-range edge, and the cap reduces how often that edge is reached. Another common combination is both VSync layers off, relying on VRR and the frame cap alone. That has lower latency, but tearing may return when the frame rate leaves the range. Neither combination is universally correct. The key is a single controlling layer, followed by verification with real gameplay, rather than assuming a menu change has taken effect.
Borderless window versus fullscreen exclusive
In fullscreen exclusive mode, the game controls display output directly, and synchronisation behaviour follows the game and driver settings closely. In borderless windowed mode, the picture passes through the system compositor, which has its own buffering and rhythm management. Tearing is usually less visible, but input lag and frame pacing may differ from fullscreen mode. Some games need an additional driver option for VRR in windowed mode; without it, VRR may not engage at all.
If tearing disappears after switching from fullscreen to borderless window, that does not necessarily mean the problem is solved. It may only mean the compositor is masking the synchronisation issue. Conversely, if borderless mode feels less responsive, switching back to fullscreen exclusive may be the more direct improvement. Always record which display mode you used when comparing settings, because the same settings can behave quite differently in the two modes.
Fast Sync and Enhanced Sync
NVIDIA Fast Sync and AMD Enhanced Sync take a different approach: the graphics card keeps drawing at full speed, but only the most recently completed whole frame is sent to the display, and older frames are discarded. This suits situations where the frame rate runs far above the refresh rate and lower VSync-style delay is desired. The graphics card still runs at full speed, and not every game benefits equally. For games whose frame rate already sits near or below the refresh rate, these technologies may bring little visible improvement and should not be treated as a universal fix.
Other easily overlooked interference
Beyond synchronisation settings, several other factors are often mistaken for graphics-card faults. Background programs such as messaging apps, browser tabs, recording software, and system updates can seize processor or storage resources at critical moments, producing short frame-time spikes. Power and performance modes matter too: on laptops, and on desktops set to a power-saving plan, the graphics card and processor may not sustain steady output. Shader compilation is another factor: when a game first enters a new scene, it may need to compile shaders on the spot, causing a brief pause that disappears on later visits to the same scene. That kind of pause is unrelated to hardware failure. These variable factors should be ruled out before considering hardware-level causes.
Part 7: Why a 144 Hz Display Can Appear Locked at 60 Hz — A Cross-Reference Checklist
This issue deserves a short separate section because it is often mistaken for a graphics-card fault, when it is frequently just an unfinished setting. The following points are a cross-reference checklist only, not the main thread of this article:
First, the operating-system display setting itself may still be at 60 Hz. Support for 144 Hz does not mean the system has selected it. The system setting and the in-game refresh-rate option are separate, and both need checking.
Second, the cable or connector may lack the bandwidth for high refresh rates. Older cables, adapters, or unsupported connectors can prevent the high refresh-rate option from appearing at all. Before replacing a cable, check the display manual for its connector requirements.
Third, the game itself may impose a limit. Some titles offer only a limited set of refresh-rate options or include a built-in frame-rate cap, making the picture look locked at a lower level. This is a game-option matter, not a display-setting matter.
Fourth, the display menu may need a relevant mode enabled first. Some displays require a high refresh-rate or overclock mode to be switched on in the on-screen display before the system offers the corresponding option. In multi-display setups, individual display settings can also affect each other and need checking one by one.
Only after all of these have been confirmed should deeper checks, such as driver versions, connector condition, and display firmware, be considered. Do not start by assuming hardware damage, because setting-level causes are far more common.
Part 8: A Step-by-Step Diagnostic Process
The principle is to change one variable at a time and verify with the same scene immediately after each change. Follow the steps in order, and do not skip ahead before earlier steps have ruled out their possibilities.
Step 1: Describe the symptom. Write down, in one sentence, what you see: a split line, uneven movement, or a late response. Note the game, the scene, and the display mode in which it appears. The more specific the description, the faster the later steps go.
Step 2: Show frame time. Most games and drivers offer an overlay that shows frame rate together with a frame-time graph. Look at the graph first, before the average number. If the graph is flat but the picture still feels wrong, the issue is more likely synchronisation or display settings. If the graph itself swings widely, the issue is more likely system load or scene-related.
Step 3: Simplify the settings. Turn off in-game VSync, set driver VSync to application-controlled or off, and disable all extra caps, returning to the simplest state. Confirm whether the basic picture is normal. Then add settings back one at a time, verifying after each addition, so you can see which setting changes the behaviour.
Step 4: Confirm that VRR is actually working. If the display has a built-in refresh-rate readout, watch whether the refresh-rate number follows the frame rate. If it stays fixed at the maximum value, VRR may not be engaging at all, and the compatibility checklist needs to be revisited item by item.
Step 5: Test one variable at a time. If background programs are suspected, close them and retest the same scene. If a multi-display setup is suspected, test with a single display first. If a particular game setting is suspected, lower it by one step and retest. Changing several things at once makes it impossible to tell which change mattered.
Step 6: Record the results. Briefly note how each setting combination behaved: which one removed tearing, which one eased stutter, which one made response feel slower. This record is very useful later, whether you seek further help or arrange an inspection, because it shows which possibilities have already been ruled out.
| Diagnostic step | What to do | If the result looks normal | If the result looks abnormal |
|---|---|---|---|
| Describe the symptom | Note the phenomenon, game, scene, and display mode | The symptom can be matched against the comparison tables, narrowing the suspects | If the description is vague, reproduce the issue once more and record it |
| Read the frame-time graph | Enable frame-rate and frame-time overlay | A flat graph points to synchronisation or display settings | A swinging graph points to load, scene, or setting conflicts |
| Simplify the settings | Switch off layered synchronisation and caps, return to basics | Basic picture is normal, so the problem lies in the setting combination | Basic picture is still abnormal, so keep checking system and game factors |
| Confirm VRR operation | Watch whether display refresh rate follows frame rate | VRR is working, so the problem may sit at the range boundary | VRR is not engaging, so revisit the compatibility checklist |
| Single-variable test | Change only one setting or environmental factor | The factor behind the change is identified | If nothing is found, move to the next variable, never several at once |
| Record the results | Write down how each combination behaved | Later comparison and help-seeking become much easier | Without notes, ineffective combinations tend to be retried |
Part 9: Recommended Starting Combinations for Different Situations
The following combinations are general starting points, not guarantees for any individual model. After applying one, always verify with a scene you play regularly, then fine-tune by feel.
Fixed-refresh display without VRR, picture integrity first: enable VSync and accept some input lag in exchange for a tear-free picture. If the frame rate often falls below the refresh rate, triple buffering can soften the effect, but watch for changes in delay.
VRR display, general gaming: enable VRR on both the display and the driver, set driver VSync on and in-game VSync off, and cap the frame rate slightly below the refresh rate. The logic is that VRR handles synchronisation inside the range, driver VSync handles the above-range edge, and the cap reduces how often that edge is reached.
VRR display, lowest latency first: enable VRR, switch both VSync layers off, and keep the frame cap slightly below the refresh rate. Latency is lower, but tearing may return when the frame rate leaves the range. This suits competitive scenes where occasional tearing is acceptable.
Frame rate far above refresh rate, lower delay desired: look into Fast Sync or Enhanced Sync style options, after confirming that the game and driver support them and that the actual picture genuinely improves. Do not assume suitability from the feature name alone.
Frame rate persistently below the bottom of the VRR range: first confirm whether the display provides LFC and what its actual range floor is. Without LFC, the picture reverts to fixed mode at low frame rates and stutter feels more pronounced. In that case, lowering game quality settings to bring the frame rate back inside the range is often more effective than adjusting synchronisation settings further.
Part 10: Setting Problem, or Something Worth a Hardware Check
Most tearing and stutter cases are setting or system-load issues, and they improve once synchronisation, caps, and game settings are put in order. A few signs, however, are worth further attention: noise, blocky artefacts, flicker, or broken textures fixed in place, appearing regardless of game or setting, even in the system interface; a display that frequently loses signal under normal settings; or a sharp deterioration in picture performance accompanied by unusually loud fans or unusual heat. These should not be addressed by settings tweaks alone, and a professional inspection is the sensible next step.
It is important to be clear that this article does not diagnose any individual model. The signs above are only prompts to seek a further check, not a diagnosis. A real judgement requires a technician to observe the actual machine, inspect connectors, and check cooling condition on site. The website's published three-tier pricing is HK$400 to HK$2,800, with the actual charge confirmed by a quotation after inspection. No repair scope is decided from a written description alone.
If you have already followed the diagnostic process and still cannot classify the problem, or if you suspect it goes beyond settings, bring your symptom notes, setting combinations, and frame-time observations when you ask for help. The address is Room 1209, Dali Plaza, Shek Mun. The telephone number is 2186 8010, and WhatsApp is 6558 6806. Complete notes help an inspection focus quickly, instead of repeating steps you have already tried.
Frequently Asked Questions
Why does the mouse feel slower after I turn on VSync, even though the picture looks smoother?
This feeling is connected to the VSync waiting mechanism. The graphics card must wait for the vertical blanking interval before swapping buffers, so the time from input to display grows longer. In fast-reaction games this delay is more noticeable; in slower-paced games it is usually less of an issue. If the delay bothers you, consider a variable refresh rate combination, or rework the frame-cap and synchronisation pairing, then verify with real gameplay.
Can a G-Sync display work with an AMD graphics card, or a FreeSync display with an NVIDIA graphics card?
Cross-brand compatibility is more common than it used to be, because many displays are based on the VESA Adaptive-Sync standard. Not every combination performs equally, however. The deciding factors remain the individual model's certification, its actual VRR range and LFC notes, and whether the driver option is correctly enabled. Check the documentation for both the display and the graphics card before assuming full compatibility from the brand name alone.
Will a lower frame cap always make the picture smoother?
No. A cap makes the rhythm predictable and helps avoid crossing the synchronisation range boundary. It is not a case of lower being better. A cap set too low directly limits how often the picture updates, making overall motion feel slower. A cap set sensibly, just below the refresh rate and inside the VRR range, supports a steady rhythm. If frame times themselves fluctuate strongly, lowering the cap alone will not fix the source of that fluctuation.
Why does the same game sometimes run smoothly and sometimes stutter, and why does restarting seem to help for a while?
On-and-off behaviour usually points to variable factors: background programs, system updates, shader compilation on first encounter, changing scene load, or power-mode switching can all make frame times worse for a stretch of time. Restarting reduces background load, so performance temporarily recovers, but the underlying cause may not be gone. When stutter appears, check the frame-time graph and background program state at that moment, rather than judging only by how things feel after a restart.
If tearing disappears in borderless windowed mode, is the problem solved?
Not necessarily. In borderless mode, the system compositor manages picture output, and its buffering can hide tearing. That does not mean the original synchronisation issue is resolved. The compositor may also add delay or change frame-pacing behaviour. Settings should be compared within the same display mode, otherwise a mode difference can easily be mistaken for a setting improvement.
My display supports 144 Hz, but the system only offers 60 Hz. Is the graphics card faulty?
Not necessarily. The most common causes are that the system display setting itself has not yet selected the higher refresh rate, that the cable or connector does not support it, that the display menu mode has not been enabled, or that the game itself imposes a limit. Work through the checklist in Part 7 first. Only after all of those have been excluded should deeper driver, connector, or hardware checks be considered. Do not assume hardware damage before the settings have been checked.
Conclusion: Name the Phenomenon Before Choosing the Fix
Tearing, stutter, ghosting, and input lag may all feel like "not smooth," but their causes and fixes are different. Tearing is a rhythm mismatch, and the fix is synchronisation. Stutter is an uneven rhythm, and the fix is frame pacing and system load. Ghosting is a display response issue, and the fix sits in the display menu. Input lag is a waiting-time issue, and the fix is less buffering and queuing. Naming the phenomenon correctly stops settings from being changed in ever more confused combinations.
A reliable starting routine is: describe the symptom, read the frame-time graph, simplify the settings and add them back one at a time, and verify every combination with the same scene. For VRR users, one principle covers most cases: let VRR lead inside the range, let a boundary mechanism handle the outside, keep the cap slightly below the top of the range, and leave only one synchronisation layer in control. With those habits, most everyday tearing and stutter can be improved substantially without replacing any hardware.