Stutter Test — Frame Time & Micro-Stutter Checker
Average FPS hides stutter; frame time exposes it. This test leads with your frame-time graph — every red bar is a frame that arrived late — plus stutter count, jitter and percentile frame times.
The meter starts automatically — no button needed. Keep this tab in front for a clean reading.
Your system (read locally, never uploaded)
Benchmark history
Why frame time, not FPS, finds stutter
Imagine two systems that both average 120 FPS. The first delivers a frame every 8.3 milliseconds like a metronome. The second delivers most frames in 6 milliseconds and then, twice a second, takes 40 milliseconds over one frame. Their averages are identical; their feel is not. The second system hitches twice a second, and you will feel every hitch, because the eye is exquisitely sensitive to rhythm breaking. FPS is a rate averaged over time, so it mathematically smooths away the very thing stutter is. Frame time — the duration of each individual frame — keeps it, which is why this page draws every frame as a bar and marks the late ones in red.
The summary statistics are chosen for the same reason. P50 is your typical frame; P95 and P99 show how bad the worst 5% and 1% get. The 1% low FPS figure converts those worst frames back into the language gamers use. Jitter compresses the whole graph into one spread number. And the stutter counter uses a plain, stated definition — any frame taking more than twice the median frame time — so a result here can be compared honestly with a re-run after you change something.
Hunting a stutter: the method
- Run a 60-second benchmark untouched and save the result with Copy Results.
- Close every other tab and background app you can, and run it again. If the red bars vanish, the cause was load — reintroduce apps one at a time.
- If stutters persist on an idle desktop, note their rhythm: evenly spaced spikes suggest a polling app or RGB utility; spikes that grow over minutes suggest heat and throttling.
- On a laptop, repeat plugged in with the performance power plan — a large share of stutter reports end there.
- Compare browsers. Stutter in one browser only points at that browser's extensions or hardware-acceleration path, not your hardware.
Stutter's neighbours are worth ruling out as you go: frames that never arrive at all are frame skipping; a trail behind moving objects is ghosting; and delay between your hand and the screen is input lag. Each has its own test here because each has a different fix, and treating one as another wastes money on the wrong upgrade.
How this test measures frame rate
Your browser draws a new frame in sync with your display, a rhythm called VSYNC, and it tells the page the exact timestamp of every frame through the requestAnimationFrame API. This tool records the gap between every pair of timestamps. Frame rate is simply 1,000 divided by the average gap in milliseconds, so the numbers here are measured, not estimated. The refresh rate is detected from the median gap, snapped to the nearest standard rate such as 60, 120 or 144 Hz. Gaps caused by switching to another tab are excluded, because browsers throttle background tabs to roughly one frame per second and that would poison the reading.
That method also defines the limit of any browser test: a page can never draw faster than your screen refreshes, so a perfect score equals your display's Hz, not an unlimited number like an in-game benchmark can report. What the browser test tells you — honestly and precisely — is whether your display, operating system settings, cable and browser are actually delivering the smoothness you paid for.
Frequently asked questions
What is stutter if my FPS is high?
Stutter is uneven frame delivery: most frames arrive on time, then one takes two or three times as long. The average FPS stays high, but that one long frame is a visible hitch. This page leads with frame time because frame time is where stutter lives.
What causes stutter?
Background work is the usual suspect — updates, sync tools, antivirus scans, other tabs — followed by thermal throttling, driver hiccups, running out of memory, and power-saving modes that keep changing the clock speed mid-test.
How many stutters are acceptable?
In a clean 60-second benchmark on an idle system: zero to a couple. A steady drip of them, or clusters at regular intervals, points at a background process on a timer — note the rhythm and hunt for what runs on that schedule.
What is jitter?
Jitter is the statistical spread of frame times — how far each frame strays from the typical gap. Low jitter at the right average means smooth; high jitter means the rhythm itself is broken even before any single frame counts as a stutter.
Can this test my games' stutter?
It tests your browser and system pipeline, which shares the CPU, GPU, drivers and background load with your games. Stutter that appears here will usually appear in games too; stutter that only appears in one game is usually that game's own streaming or shader compilation.