Network Jitter Test
AvailableTest how stable your connection is: repeated round-trip measurements from your browser with jitter, latency, failed probes, live charts and a stability report.
Browser-measured latency. The default mode runs in your browser and makes real measurements of HTTP round trips to the chosen server. It is not ICMP ping. Server ICMP sends real ICMP echo requests from the NetTools server to a host (10 or 30 s tests). Server UDP and WebRTC are not available yet.
Test settings
ReadyJitter report
No test yet
Live results appear here while the test runs.
- Jitter
- —
- Average latency
- —
- Min / Max
- —milliseconds
- Failed probes
- —
- Probes
- —
- Stability
- —
Live charts
Results
Latency
Jitter
Probes
Test
Stability summary and suggestions
How these numbers were measured
- Jitter = average(|RTT[i] − RTT[i−1]|) over consecutive probes that both succeeded. A failed probe breaks the sequence and is never counted as 0 ms.
- This is a simple variation metric. It is not RTP interarrival jitter (RFC 3550), which VoIP equipment reports as a smoothed one-way estimate, and other tools may use standard deviation instead. Compare numbers measured the same way.
- Browser mode measures HTTP round trips: request to first byte of the response where the browser exposes it (Resource Timing), otherwise the elapsed request time. Both include a little browser and server overhead and are usually a few ms above ICMP ping.
- One warm-up request opens the connection first and is not counted. A probe without a response within 2 s counts as timed out.
- Stability rating: very stable ≤ 5 ms jitter and no failures; stable ≤ 15 ms and ≤ 1% failed; somewhat unstable ≤ 30 ms and ≤ 3%; unstable above. Rules of thumb for calls and games, not a standard.
How it works
Your browser sends a small, uncached HTTP request at a fixed interval and times each round trip. Jitter is the average change between consecutive successful measurements: average(|RTT[i] − RTT[i−1]|). Failed requests are counted separately and never as 0 ms. These are browser-measured HTTP round trips, not ICMP ping, so they include a little browser and HTTP overhead.
Reading the results
- Latency is how long a round trip takes. Jitter is how much that time changes from one probe to the next. Calls and games suffer more from high jitter than from steady high latency.
- Test a few times and, if possible, on a wired connection for comparison. Wi-Fi is the most common source of jitter at home.
- The test pauses while this tab is in the background (browsers slow down background timers); a pause longer than 15 seconds ends the test as interrupted.
Frequently asked questions
What is a good jitter value?
Under about 5 ms is excellent, under 15 ms is fine for nearly everything, and up to about 30 ms is usually acceptable for voice and video calls. Above that, calls and games tend to stutter.
How is jitter calculated here?
As the average absolute difference between consecutive round-trip times, using only pairs where both probes succeeded. Other tools may use other definitions (standard deviation, or the smoothed RFC 3550 interarrival jitter used by VoIP equipment), so numbers are not always directly comparable.
Is this the same as ping?
No. Ping uses ICMP packets sent by the operating system. This test uses HTTP requests from your browser, which works without installing anything but adds some overhead. The results describe your real connection to the chosen server.
Does a failed probe mean packet loss?
Not necessarily. A browser request can fail for reasons other than lost packets: an extension, a rate limit, the server or the browser itself. Repeated failures are a reason to confirm with ping or MTR.
Why does the test pause when I switch tabs?
Browsers slow down timers in background tabs, which would distort the measurements. The test pauses while the tab is hidden and continues when you come back; long pauses end the test as interrupted.