What is jitter? A plain-English guide to good and high jitter

As a rule of thumb, under 5 ms is excellent, under 15 ms is good for nearly everything, and above 30 ms you will notice problems in live audio and video. The right target depends on what you are doing, so here are both views: a simple scale, then a breakdown by activity.

Jitter, latency and packet loss travel together, so here is the same idea by activity. The Teams and Zoom rows use NetCheckPro’s readiness thresholds, the VoIP row comes from Cisco, and gaming and streaming are rules of thumb because no vendor publishes a standard for them.

These terms get mixed up constantly. Latency is the delay: how long a packet takes to get from you to a server. Ping is the everyday name for a round-trip latency test, reported in ms. Jitter is how much that delay changes from one packet to the next. Packet loss is a separate problem, where packets never arrive at all.

Think of a commute. Latency is the average journey time. Jitter is how unpredictable it is: a 20-minute drive that sometimes takes 45 is worse for catching a train than a steady 30. A connection can have low ping and high jitter, or high ping and almost no jitter. Both matter for calls, but jitter is what makes audio sound broken. Use the latency test for delay, the jitter test for consistency, and the packet loss test for dropped packets. For a side-by-side comparison, read jitter vs latency vs packet loss.

High jitter is a sustained average above 30 ms for anything real-time, or above 15 ms for fast games and clean phone-quality audio. One spike does not count; repeated large swings do. Typical symptoms:

• Voices that sound robotic, or words that clip at the start and end. Microsoft ties robotic-sounding speakers to jitter on Teams calls (source).
• Video that freezes, then jumps forward to catch up.
• Games with rubber-banding, delayed hits or lag spikes while the ping counter looks normal.
• Stuttering screen shares and inconsistent VPN sessions.

If your result is high, jump to how to fix it or read our step-by-step high jitter fix guide.

It depends on whether the app can wait. Web pages, downloads and email retransmit and reorder data, so they barely notice jitter. Live audio, video and games cannot wait.

• Video calls and VoIP: aim for under 30 ms and preferably under 15 ms.
• Competitive gaming: under 5 ms feels best; up to about 15 ms is playable.
• Streaming and browsing: 30 ms or more is usually fine.

Vendor guidance frames the range. Cisco’s Webex Calling documentation gives 10 ms as the ceiling for its Local Gateway link, while Microsoft only marks a Teams audio stream as poor once average jitter passes 30 ms. Apps absorb some jitter with a buffer, so the answer is a range. If a 20 ms result sounds fine on your calls, it is fine.

Jitter happens when packets wait in different queues for different lengths of time. The usual culprits, roughly in order of how often we see them:

• Wi-Fi interference and weak signal. Radio is shared and retries lost frames, so timing varies. Microsoft notes 2.4 GHz is often disturbed by other consumer devices, while 5 GHz suits real-time media better (source). Our Wi-Fi stability test helps isolate this.
• Bufferbloat. When your router’s queue fills during an upload or download, your call’s packets wait behind the bulk traffic. Check with the bufferbloat test.
• Congestion. Busy evenings on shared links add unpredictable waits.
• VPNs. Extra encryption, and traffic hairpinned through a distant server, can add latency and jitter; Microsoft recommends split tunnelling for Teams for this reason (source).
• ISP routing. A longer or more congested path gives delay more places to vary.
• Overloaded devices. Old routers, cheap extenders, a maxed-out CPU or a phone backing up photos can all disrupt timing.

NetCheckPro takes 26 idle latency samples from your browser, timing round trips to a nearby Cloudflare server before it loads the connection. Jitter is the average difference between each sample and the one before it, which is the same calculation as the worked example above. It runs in your browser, with nothing to install.

NetCheckPro does not measure packet loss, because a browser cannot send the raw packets that measuring loss needs, and the result says “not measured” instead of guessing. Treat jitter as a good estimate rather than a lab reading: it times web requests, not the UDP media stream a call uses, and RFC 3550 defines a smoothed version that differs slightly. For best results:

• Test on the device and network you actually use, ideally while the problem is happening.
• Repeat at different times of day, and once over Ethernet, to separate line problems from Wi-Fi problems.
• Compare with your call app’s own numbers. Zoom’s statistics tab, for example, reports latency, jitter and packet loss during a meeting.

Work through these in order. The first three fix most home problems.

1. Test on Ethernet. Plug a laptop straight into the router and run the test again. If jitter drops, the problem is your Wi-Fi. If it stays high, the router or the line is responsible.
2. Use Ethernet for calls and games. A cable removes radio interference entirely. Powerline or MoCA adapters beat a weak Wi-Fi link if you cannot run one.
3. Fix your Wi-Fi. Switch to the 5 GHz or 6 GHz band, move closer to the router, and pick a less crowded channel. Avoid relying on range extenders for calls.
4. Turn on QoS or SQM to fix bufferbloat. Smart Queue Management (fq_codel or CAKE) or your router’s QoS setting keeps queues short. Set the limits slightly below your measured line speed, then re-run the bufferbloat test.
5. Pause heavy uploads and downloads. Cloud backups, game updates and 4K streams are common triggers. See low upload speed fix if uploads swamp your line.
6. Update or replace the router. Restart it, install firmware updates, and if it is five or more years old, budget for a modern Wi-Fi 6 or Wi-Fi 7 model.
7. Turn off the VPN, or split-tunnel your calls. Compare with and without using the VPN on vs off test.
8. Contact your ISP with evidence. Send screenshots of results taken on Ethernet at several times of day and note the worst times. That gives support something to escalate.

Call and game software fights jitter with a jitter buffer. Instead of playing each packet the moment it arrives, the app holds a few milliseconds of audio and replays it at a steady pace. A late packet that still fits inside the buffer is never heard.

There is a trade-off. Microsoft notes that buffering for jitter increases end-to-end latency (source), so a larger buffer means a longer pause before people hear each other. Cisco adds that jitter buffers are usually only effective on variations below 100 ms (source), and a buffer that is too small leaves gaps in speech (source). Apps size the buffer automatically, so fix jitter at the source and let the buffer handle only what is left.