Media diagnostics · Primary-source explanation; no UK IPTV implementation or account test. Instructions are based on publisher guidance; a UK IPTV account was not tested on this platform.
One hundred invented observations
Suppose ninety-nine packet delay observations are ten milliseconds each, and the remaining observation is 1,010 milliseconds. Their total is 2,000 milliseconds, so the mean is twenty milliseconds. The median is ten milliseconds and the largest observation is 1,010 milliseconds. All three are correct summaries of the same data. Describing the session only as twenty milliseconds discards the unusual event that may be most relevant to a report about one noticeable interruption.
Choose and name the statistic
RFC 5481 discusses delay variation formulations and the role of summary statistics in their intended uses. It does not make every quantity labelled jitter equivalent. The numbers in this example are raw hypothetical delay observations, not a computed RTCP jitter field. A percentile can be useful for a larger measurement, but its sample definition and calculation method should accompany it. Do not assign a precise percentile label to this small example without stating the method, especially near the boundary between the common values and the single largest observation.
Preserve the event context
Record the observation count, collection interval, units, endpoint and whether loss or timeouts were excluded. Keep the time of the unusual observation so that it can be compared with the reported symptom. A large network delay observation is still not direct proof of a one-second picture freeze: buffering may absorb it, and a freeze can have other causes. The useful decision is to investigate the tail event alongside the average, rather than allow a compact dashboard number to replace the original evidence.