Network Jitter Test

Two connections can have the same average delay and feel completely different. One is steady; the other fluctuates between fast and slow. Voice calls, video meetings and games notice the fluctuation much more than the average. The page measures both from your browser and presents the numbers in a way you can compare between runs, networks and times of day.

Latency and jitter

Latency is the time a reply takes to come back, here measured as the round trip of a small HTTPS request. Jitter is how much that time varies. The page defines jitter as the mean absolute difference between consecutive round-trip times, a simple version of the interarrival jitter idea used in RTP (RFC 3550). It also shows the standard deviation, the lowest and highest values, the median and the 95th percentile, which is the value below which 95 percent of the replies fell.

Running the test

Choose the number of requests, between 5 and 200, and the pause between them in milliseconds, then start it. Each request goes to this site (openworkstand.com), so the test measures the path between your browser and the edge of the site’s network, not the whole Internet and not a particular game or conferencing server. Nothing else is contacted and no personal data is sent. Lost or timed-out requests are counted. A line chart shows each reply, so you can see spikes at a glance, for example when another download starts.

Reading the verdict

The page grades the run, for instance “Stable”, and explains it in words. As its guidance says, for voice and video aim for jitter under 30 ms and almost no loss; for games a low and steady average matters most. Peaks in the chart show moments when a wireless link, a busy router or another download delayed your traffic. Run the test several times, on a wired connection and on Wi-Fi, and while nothing else is using the line, to see which factor matters.

A worked example of the jitter figure

Suppose five replies take 20, 24, 21, 30 and 22 ms. The differences between neighbours are 4, 3, 9 and 8 ms, and their mean is 6 ms, so the jitter is 6 ms. The average time is 23.4 ms and the median is 22 ms. A second connection with replies of 23, 23, 24, 23 and 24 ms has the same average but a jitter of 0.75 ms. The first would struggle with a call and the second would not. These figures are an illustration of the calculation, not a measurement.

Limits

Browsers cannot send ICMP pings, so each time includes some work by the server and the browser, and the first request may include opening a connection. Timer resolution in browsers is deliberately coarse, which affects very small differences. The test shows how your connection to this site behaves now; the route to another service may differ. For the effect of delay on download speed see the TCP Window Calculator, and for call bandwidth the VoIP Bandwidth Calculator.

Related tools

• VoIP Bandwidth Calculator sizes the link for calls whose quality depends on jitter.
• TCP Window Calculator shows how round-trip time limits single-connection speed.
• HTTP/3 Test reports which protocol your browser uses for this site.
• Wi-Fi Signal Calculator interprets signal readings when Wi-Fi is the weak link.