This stutter test measures how evenly your frames arrive, not just how many of them there are. A display can report a high average frame rate and still feel rough, because a single frame that arrives late — or never arrives — breaks the rhythm of motion. That hitch is stutter, also called jank, and your eyes notice it even when the average number looks perfect. The test moves a marker at a fixed step every frame and plots the real time between frames underneath, turning a feeling into a number you can compare.
Smooth motion depends on consistent frame timing. At 60 FPS a new frame should appear every 16.7 milliseconds; at 120 FPS, every 8.3. When one frame takes two or three times that long, the motion jumps forward, and no amount of high average frame rate hides it.
The test starts automatically. There are two things to watch:
The cards summarise the run. Expected frame time is your typical gap between frames; Worst frame gap is the longest single delay recorded; and Janks counts every late frame. Use Reset meter to clear the totals, then leave the test running for a while, or switch tabs and come back, to see how your system behaves over time.
Stutter is a symptom of the rendering pipeline missing its deadline. The frame is late leaving your computer, before the display is ever involved. Common causes include:
Another browser tab, a software update, a virus scan, or a sync client can steal CPU time for a moment and push a frame past its deadline. Stutter that comes in bursts, with long quiet stretches between, usually points to something else on the system waking up.
When the graphics card or processor is fully loaded, some frames take longer than others to build. This shows on the graph as a rough, uneven band rather than isolated spikes. Closing heavy applications or lowering a game's settings smooths it out.
Laptops in a battery-saver mode, or any device that has grown hot, deliberately slow down. The frame rate sags and frame times wander. Plugging in, improving airflow, or switching to a performance power plan often steadies the timing immediately.
Layout recalculation, garbage collection, and heavy scripts can each pause rendering for a frame or two. This is why the same page can feel smooth in one browser and janky in another on identical hardware.
This is the heart of why the test exists. Consider one second that contains 118 frames of 8 milliseconds and two frames of 25 milliseconds. The average is still roughly 120 FPS — a number that looks flawless — yet those two long frames are visible hitches. A frame-time view catches them; an averaged frame counter never will.
| Metric | What it tells you | Blind spot |
|---|---|---|
| Average FPS | Overall throughput | Hides isolated late frames |
| Frame time (per frame) | The rhythm of motion | Needs a graph to read at a glance |
| Jank count | How often motion breaks | Does not say how severe each hitch is |
| Worst frame gap | The severity of the worst hitch | A single event, not the trend |
Reading all four together gives the full picture. A good result is a high, steady frame rate with a low jank count and a worst gap close to the expected frame time. You can cross-check the raw frame rate on the main FPS test, and our overview of frames per second explains why the average alone is not enough.
Once the meter shows a problem, work through the likely causes:
Because average frame rate and frame consistency are different things. A few late frames barely move the average but are clearly visible and are counted here as janks. That is exactly the case this test is built to expose.
An occasional isolated spike, especially right after the page loads or when you switch tabs, is normal. A steady stream of red bars, or a worst gap several times your expected frame time, points to a real problem worth chasing.
Not necessarily. Background activity, thermal throttling, power saving, and browser hitches all cause stutter on capable hardware. Work through those before assuming you need an upgrade.
Stutter is a timing problem — frames arriving late — so the whole image hitches but stays intact. Tearing is a synchronisation problem where a frame is split across a refresh. See the screen tearing test for that one.
There is no single pass mark, because it depends on how long you run the test, but the shape of a healthy result is clear: the jank count should climb slowly or not at all once the page has settled, and the worst frame gap should stay within roughly one and a half times your expected frame time. A jank count that ticks up several times a second, or a worst gap several times the expected figure, means motion is visibly breaking and one of the causes above is at work.
Reviewers rarely judge smoothness by average frame rate alone. They also report 1% and 0.1% lows — the frame rate of the slowest one per cent and one per thousand of frames. These figures describe the same thing this test measures: the worst moments, not the average. A game with a high average but poor 1% lows feels rough, exactly as a display with a high frame rate but frequent janks does here. The jank count and worst frame gap on this page are the browser equivalent of those percentile lows.
Micro-stutter is a subtler cousin. Instead of dramatic spikes, frame times alternate — long, short, long, short — so the average looks fine and even the worst gap is modest, yet motion feels slightly uneven. On the frame-time graph it appears as a jagged band rather than tall isolated spikes. Micro-stutter usually comes from imperfect frame pacing, where frames are produced faster than they can be shown evenly, and it is one reason a locked, well-paced frame rate can look smoother than a higher but erratic one.
It seems backwards, but limiting your frame rate often improves how motion feels. An uncapped frame rate that swings between, say, 90 and 200 constantly changes the frame time, and that variation is visible. Capping it to a steady value your system can always hit keeps every frame the same length, which is what the eye reads as smooth. If this test shows a high average but a restless graph, try capping the source that feeds it and watch the band settle — a steady 100 frames per second beats an average of 140 that lurches. Matching your frame rate to your refresh rate follows the same logic, and a variable refresh rate helps most by moving the refresh to meet each frame rather than forcing frames to wait.
A single snapshot can mislead. Leave the test running and watch for patterns: a burst of red every few seconds suggests a background task on a timer, a slow drift toward longer frame times hints at thermal throttling, and a permanently rough band points to a system under constant load. Resetting the meter and comparing a quiet minute against a busy one — with other tabs open, or a download running — turns the test into a simple diagnostic for what on your machine is stealing frames.
Frame consistency is one part of smooth motion. Continue with these companion tools: