Buffering

The spinning-wheel pause that kills a viewing session — and the problem most of streaming's core technology exists to prevent.

Updated: May 2026 4 min read Streaming
In short

Buffering is when a video stalls mid-playback because data arrives slower than it plays; the player's reserve of upcoming footage empties, so it pauses to reload before continuing.

What is buffering?

Buffering is when video playback pauses to download more data before it can continue. The player keeps a small reserve of upcoming video (the "buffer"); when data arrives slower than it's played, that reserve empties and playback stalls — the spinning wheel — until enough has loaded to resume.

It usually comes down to one of a few causes: the bitrateBitrateThe amount of data used per second of video. Higher bitrate means more detail and a larger file; the core lever behind streaming quality.Open definition → is too high for the viewer's connection, the server or CDNCDN (Content Delivery Network)A global network of edge servers that caches video close to viewers, cutting latency and absorbing traffic spikes.Open definition → is far away or overloaded, or the video wasn't prepared for adaptiveAdaptive Bitrate (ABR)Automatically adjusts video quality in real time to match each viewer’s connection speed — smooth playback from 480p to 4K with no buffering.Open definition → delivery. Rebuffering is one of the strongest predictors of viewers abandoning a video, which is why so much of streaming's technology exists to avoid it.

Initial buffering vs rebuffering

Not all buffering is equal, and the industry measures the two kinds separately as quality-of-experience metrics.

TypeWhen it happensMetricImpact
Initial bufferingBefore playback startsVideo start timeViewers abandon before watching
Rebuffering Mid-playback stallsRebuffer ratio (% of watch time stalled)Most damaging — interrupts committed viewers

A low rebuffer ratio and a fast start time are the two numbers that best predict watch-through — worth tracking in your analytics.

How buffering is prevented

Adaptive bitrate

The player steps quality down before the buffer empties, so playback continues at a lower resolutionResolutionThe pixel dimensions of a video frame — 480p, 720p, 1080p, 2160p (4K). Determines how sharp the picture looks on a given screen.Open definition → instead of stalling — then steps back up when the connection recovers.

A fast CDN

Serving segments from an edge serverCDN (Content Delivery Network)A global network of edge servers that caches video close to viewers, cutting latency and absorbing traffic spikes.Open definition → near the viewer keeps data arriving quickly and reliably, keeping the buffer full even under load.

Sensible bitrate targets

A well-built rendition ladder gives the player realistic rungs to pick from, so it can always find one the connection can sustain.

A healthy buffer window

Holding a few seconds of video in reserve lets the player ride out short bandwidth dips without stalling — a trade against latencyLatencyThe delay between a moment happening and a viewer seeing it. Critical for live — Low-Latency HLS brings it down to 2–5 seconds.Open definition → on live.

Further reading

Denis Konnov

Written by

Denis Konnov Verified

CMO at Kinescope

Denis leads marketing and content at Kinescope, working closely with the platform's engineers. This glossary is reviewed and kept up to date by the team building Kinescope's streaming, encoding and CDN delivery infrastructure.

FAQ

Can’t find the answer you need? Check out our Help Center.

Buffering happens when the player can’t download data fast enough to keep playing — usually because the bitrate is too high for the connection, the server (or CDN) is far away or overloaded, or the video wasn’t prepared for adaptive streaming. Adaptive bitrate fixes most of it by stepping quality down before the buffer empties.

Buffering is playback pausing to load more data mid-stream. Latency is the delay between something happening (or you pressing play) and seeing it. A live stream can have low latency but still buffer on a weak connection, and vice versa — they are different problems.

Initial buffering is the wait before playback starts (measured as video start time). Rebuffering is a stall in the middle of playback after it has already begun. Both hurt, but rebuffering is the more damaging: a viewer who has committed to watching and then gets interrupted is far more likely to abandon.

Adaptive streaming encodes the video at several quality levels and lets the player drop to a lower one when bandwidth falls, so playback continues smoothly at reduced quality instead of stalling. It steps quality down before the buffer runs dry, then back up when the connection recovers.

Yes. A CDN serves video from an edge server near the viewer, so segments arrive faster and more reliably — which keeps the buffer full and reduces stalls, especially under load or for distant viewers. It’s one of the biggest levers on rebuffering at scale.

Not entirely — a viewer with almost no connection will eventually stall. But you can make it rare: adaptive bitrate, a fast CDN, sensible bitrate targets and a healthy buffer window together keep rebuffering close to zero for the vast majority of viewers.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.