The short definitions
- Latency is the delay: how long one packet takes to get to a server and, for a ping, back again. It is measured in milliseconds (ms). Everyday “ping” is a round-trip latency reading.
- Jitter is the variation in that delay from one packet to the next, also in ms. Steady delay means low jitter, and delay that bounces around means high jitter.
- Packet loss is the share of packets that never arrive, shown as a percentage. The packet was sent, but it was dropped on the way.
A fourth number people mix up with these is speed, or throughput. That is how much data fits through per second, in Mbps, and it is independent of all three. Our what is jitter guide goes deeper on the middle one.
Jitter vs latency vs packet loss: comparison table
| Feature | Latency | Jitter | Packet loss |
|---|---|---|---|
| What it is | The time a packet takes to travel, or a round trip. | How much the delay changes between packets. | Packets that are sent but never arrive. |
| Unit | Milliseconds (ms) | Milliseconds (ms) | Percent (%) of packets lost |
| What you notice | Pauses before replies, people talking over each other, sluggish games and remote desktops. | Robotic or stuttering audio, video that freezes then jumps forward, rubber-banding in games. | Missing words, choppy or frozen video, dropped game actions, stalled VPNs. |
| Common causes | Distance to the server, long routes, VPNs, a router queue that fills under load. | Wi-Fi interference, congested links, bufferbloat, an overloaded router or device. | Weak Wi-Fi, a congested or faulty link, bad cables, an overloaded device, equipment that drops packets when overwhelmed. |
| How to check it | Ping, or the NetCheckPro latency test. | The NetCheckPro jitter test, or the jitter figure in your call app. | Not in NetCheckPro. Use call-app statistics or 100 pings (see below). |
| Best first fix | Use a nearer server, turn off the VPN, control bufferbloat. | Use Ethernet, improve Wi-Fi, turn on SQM or QoS. | Test on Ethernet, replace faulty cables, escalate to your ISP. See how to fix packet loss. |
Typical symptoms and causes, not a diagnosis. Several problems can share a symptom, which is why you test the measurements separately.
An analogy: reading a letter by post
Imagine a long letter, posted one page at a time. Latency is how long a page takes to reach you. Jitter is whether the pages arrive evenly or in unpredictable clumps: a page a day is easy to follow, but a week of nothing followed by five pages at once is not. Packet loss is pages that never turn up.
If you have to read the story aloud to a friend as it arrives, as in a live call, long delivery times make the conversation slow. Uneven arrival forces you to stop and wait, and a missing page leaves a gap that nobody can fill in time. If you are only collecting a package to read later, such as a download, none of that matters much, and only the total time does.
Which one matters for which activity
| Activity | Matters most | Why |
|---|---|---|
| Video calls and VoIP | Jitter and packet loss, then latency | Audio and video arrive as a live stream. Late or missing packets cannot be replayed, and high latency makes turn-taking awkward. |
| Online gaming | Latency and jitter, then packet loss | Games send small updates many times a second. Steady, low delay feels responsive, and lost updates cause rubber-banding. |
| Streaming video | Throughput first; the rest matter less | Players buffer ahead, so they ride out jitter and small delays. Heavy loss can still cause stalls. |
| Web browsing and downloads | Throughput and latency | Pages and files are retransmitted if a packet is lost, so loss mostly shows as slower loading rather than gaps. |
| Remote desktop and VPN sessions | Latency, then jitter | Every keystroke and click waits for a reply, so delay and uneven delay feel like lag. |
General guidance based on how each app works. For benchmarks by app, see the Teams and Zoom pages linked below.
For specific numbers, the Teams speed test, Zoom speed test and gaming ping test pages show targets for each use.
What vendors consider good
No single standard exists, and the figures below differ because each vendor measures in its own way. They are the published figures we could verify, shown side by side.
| Source | Latency / round trip | Jitter | Packet loss |
|---|---|---|---|
| Teams Call health, typical (Microsoft) | Under 200 ms | Under 30 ms (received audio) | Under 2% (received) |
| Zoom Statistics tab, recommended (Zoom) | 150 ms or less | 40 ms or less | 2% or less |
| Teams call-quality dashboard, poor stream above (Microsoft) | Over 500 ms round trip | Over 30 ms average | Over 10% |
The last row is a failure line, not a target. Microsoft notes that a stream flagged as poor does not always mean users noticed.
How latency, jitter and packet loss affect each other
The three are measured separately, but they often share a cause, which is why one problem tends to bring the others.
- A full router queue raises latency and jitter together. Packets wait behind bulk traffic, and the wait is uneven. Under a big upload, ping rises and wobbles. See why ping is high when someone is streaming or uploading.
- Jitter buffers trade jitter for latency. Call apps hold a little audio and replay it at a steady pace. Microsoft notes that buffering for jitter increases latency (archived Microsoft page), so smoothing is never free.
- Late can be as bad as lost. A packet that arrives after its moment in a live call has no use, so extreme jitter can sound like packet loss.
- Congestion can cause loss and delay at once. A link that is overloaded can queue packets, and drop some when the queue is full. Bufferbloat.net explains that oversized buffers were meant to avoid drops, but pay for it in delay (source).
- Wi-Fi retries add both. A radio frame that is lost is sent again, which delays it and makes timing uneven, and if retries run out the packet is lost.
That is why the first step in nearly every fix is the same: test on Ethernet. If all three improve, the cause was Wi-Fi.
What NetCheckPro measures, and what it does not
- Latency: the median of 26 idle round trips to a nearby Cloudflare server.
- Jitter: the average difference between consecutive idle latency samples. RFC 3550 defines a smoothed version used by call software, so the figures are close in spirit but not identical.
- Bufferbloat: the extra latency while the line is busy with downloads and uploads.
- Packet loss: not measured.
NetCheckPro does not measure packet loss. A browser cannot send the raw packets that measuring loss needs, and a web request is retried automatically if a packet goes missing, so loss shows up as delay, not as a gap. The result says “not measured” rather than guess. Our how we measure page explains the method and its limits.
It also times web requests, not the UDP stream a call uses, so treat it as a strong hint about your connection and not a verdict on one call. Clean latency and jitter rule out many common causes of choppy calls, but only a real loss measurement can rule loss in or out.
How to check packet loss properly
- Use your call app while the problem is happening. In Teams, open the meeting controls and choose More actions, then Settings, then Call health, to see received packet loss, round-trip time and received jitter (Microsoft). In Zoom, open the Statistics tab from Video Settings, which shows latency, jitter and packet loss for audio, video and screen sharing (Zoom).
- Send 100 pings. On Windows run
ping -n 100 1.1.1.1. On macOS or Linux runping -c 100 1.1.1.1. Read the loss figure in the summary. Our packet loss test page walks through it, including finding which hop is dropping packets. - Repeat over Ethernet and Wi-Fi. A difference points to your wireless link. Loss on a cable points to the router, the cable or the line.
- Check your router and your provider. Some routers and modems show error or dropped-packet counters on their status pages, and your ISP may be able to test the line. Take your numbers and times with you.
- Treat one result as a hint. Some routers deprioritise ping replies, so repeat the run a few times.
If you find loss, follow how to fix packet loss. If latency and jitter look high instead, see how to fix high jitter.
Sources
Figures attributed to a vendor come from the pages below. Thresholds marked as NetCheckPro guidelines are our own rules of thumb, not an industry standard.
- Monitor call and meeting quality in Microsoft Teams – Microsoft Support. How to open Call health, and typical values for round-trip time (under 200 ms), received packet loss (under 2%) and received jitter (under 30 ms).
- Accessing meeting and phone statistics (Zoom) – Zoom Support. How to open the Statistics tab, and recommended latency of 150 ms or less, jitter of 40 ms or less, packet loss of 2% or less.
- Use CQD to manage call and meeting quality in Microsoft Teams – Microsoft Learn. Poor-stream thresholds: 30 ms jitter, 10% loss, 500 ms round trip, and what each sounds like.
- Media Quality and Network Connectivity Performance (archived) – Microsoft Learn (retired page). Jitter definition and why buffering for jitter adds latency.
- RFC 3550: RTP, a transport protocol for real-time applications – IETF. Formal definition of interarrival jitter.
- Bufferbloat: introduction – Bufferbloat.net. Oversized buffers add latency, and the internet’s protocols are designed to tolerate some packet loss.