Buffer fill (rebuffer duration)
Buffer fill is the total time, in seconds, that a viewer spent waiting while the player loaded more video after playback had already started. It is the raw number behind buffer ratio. A 30 second buffer fill on a 5 minute video is a bad experience; the same 30 seconds across a 3 hour match is barely noticed.
Why it matters
Buffer fill turns a percentage into seconds that anyone in the business can grasp. Saying viewers lost 40 seconds to loading during a final makes a delivery problem concrete, and it helps justify spending on CDN capacity or encoding changes. If those lost seconds point to delivery, teams streaming on FastPix already have multi-CDN delivery, so a second CDN is not a separate purchase to argue for.
How it works
- Each time playback stalls after starting, the player records when the stall began.
- When playback resumes, it records the end time and works out the stall's length.
- All stall lengths in the view are added up to give buffer fill in seconds.
- Dividing buffer fill by watch time gives the buffer ratio.
Is buffer fill the same as rebuffer duration?
Yes, buffer fill and rebuffer duration are two names for the same measure: total seconds spent stalled after playback began. Some tools also report the average length of a single stall, which is a different number. Check that startup waiting is excluded. It normally is, because that belongs to startup time.
How much buffering is acceptable?
There is no fixed limit, because acceptable buffering depends on how long the view is. A few seconds across a two hour film may go unnoticed, while the same few seconds in a 30 second clip ruins it. That is why teams judge buffer fill alongside watch time, using buffer ratio as the comparable figure.
Where it's used
Live event platforms and OTT services use buffer fill to put a wall-clock figure on delivery problems, which is easier to explain to a business team than a percentage.
Where it usually goes wrong, and how FastPix handles it
Buffer fill on its own looks worse on long views and better on short ones, so teams that compare raw seconds across content types draw the wrong conclusions.
Buffer fill is the sum of all buffering events in a view in FastPix Video Data, and dividing it by watch time gives buffer ratio. Lost seconds can be quoted to the business while still comparing fairly across short clips and long films.
