IsMyInternetStableTest · Diagnose · Fix

    Video Call Stability Test

    This test runs a small simulated call: a stream of video comes down and a stream goes up, the way they do in a real call, while it watches your delay and jitter. Every second that a stream is late is a moment a real call would freeze, and the result is an estimated call quality score from 1 to 5.

    Test duration

    Short tests are useful for quick checks. Longer tests are better for finding intermittent problems.

    Free, no signup. Sends call-like traffic both ways: about 15 MB in total for this duration, so avoid it on metered data.

    Customize: goal, connection type

    Connected via (optional — sharpens the diagnosis)

    What the call simulation measures

    Their video, coming in

    A steady 2.5 Mbps stream arrives once a second, like the other person on a call. We report how much of it arrived in time.

    Your video, going out

    A steady 1.5 Mbps stream leaves once a second, like your own camera. Upload is the usual weak side of home connections, so it is scored separately.

    Freezes

    A stream chunk that is not complete within its second is a moment a call would stall or break up. Neighbouring late seconds count as one freeze, and the longest is reported.

    Estimated call quality

    Delay, jitter and late chunks are combined into a score from 1 to 5, the way phone-call quality is rated (MOS). A clean line tops out around 4.4, so a steady 4.4 means nothing is holding the call back.

    What the animation is

    The two talking-head tiles are an illustration: they blur, pixelate and freeze when the measured jitter, lost requests and disconnects say a real call would. The numbers around them are the measurement.

    Why calls are more sensitive to jitter and loss than raw speed

    A video call is a small, steady stream that has to arrive on time, every time. A fast connection with occasional jitter or loss can feel worse on a call than a slower, steadier one, because the call app has very little time to wait for a late packet.

    That is why this test sends call-like traffic in both directions instead of only timing small requests, and why it reports late seconds and freezes rather than just an average.

    Frequently asked questions

    No. It is a simulation: two steady streams of call-like traffic, one each way, with your delay and jitter measured at the same time. The talking-head tiles are an illustration driven by the same measurements.

    It estimates how a call would sound and look from your delay, jitter and late chunks, on the 1 to 5 scale used for phone calls (MOS). Above about 4.0 is good. A clean connection tops out near 4.4, so a steady 4.4 simply means nothing is hurting the call.

    On most home connections upload is much slower than download, and it is the first thing to struggle when something else is uploading. If only your own video freezes, the other person sees you stall while you still see them fine.

    Most often a brief disconnect or a run of lost packets. Jitter causes choppy audio more than full freezes. Run this test to see whether any of those happened during the window you tested.

    Test on whichever you actually use for calls, then try the other — the Wi-Fi stability test has a guided Wi-Fi vs. Ethernet comparison if you want to isolate the cause.

    Related