Rebuffering

Rebuffering refers to a pause in streaming playback caused by the player running out of buffered data and waiting for additional data to arrive before continuing. It is distinct from the initial buffering that may occur before playback starts or resumes after a seek. In adaptive streaming, the client maintains a buffer of several seconds of content, and rebuffering occurs when the data arrival rate cannot keep up with the playback rate, causing the buffer to deplete.

Rebuffering events are typically described by two metrics: rebuffering time and rebuffering frequency. Rebuffering time is the total duration of all stalls during a viewing or listening session, while rebuffering frequency is the number of stalls over a given period. Together, these metrics reflect the user experience and quality of service.

Causes of rebuffering include fluctuating network conditions, high or unstable throughput, latency in data delivery, server or CDN load, and aggressive bitrate switching in ABR (adaptive bitrate) strategies. Long startup delays after seeking or network handoffs between networks can also trigger rebuffering. Content with very high bitrates relative to available bandwidth is more prone to rebuffering, as is scenarios with bursty traffic or poor wireless connectivity.

Impact on user experience can be significant, often leading to increased drop-off, reduced engagement, and negative perceptions of service quality. To mitigate rebuffering, streaming systems employ techniques such as adaptive bitrate algorithms, larger initial buffers, smoother bitrate transitions, prefetching and caching, shorter segment durations, and improved network delivery through CDNs and edge servers.