JitterBuffer

A jitter buffer is a buffer used in real-time communications to counteract jitter, the variation in packet arrival times caused by network conditions. It stores incoming packets for a short, controlled period and releases them at regular intervals to produce smooth, continuous playback for audio or video streams.

Operation typically relies on sequence numbers and timestamps. As packets arrive, they are placed in the buffer and scheduled for playout according to their timing information. If a packet arrives late, it may be discarded or used for packet loss concealment. The buffer may also reorder out-of-sequence packets to preserve correct playback order. Some implementations adjust their behavior based on observed network delay statistics.

Types of jitter buffers include fixed-size and adaptive variants. A fixed-size buffer provides predictable latency but can increase dropout under high jitter. An adaptive jitter buffer changes its size in response to observed jitter, aiming to balance latency and packet loss. Adaptive algorithms may use moving averages of inter-arrival times, thresholds, or control-loop techniques to determine when to grow or shrink the buffer.

Trade-offs are central to jitter buffer design. Larger buffers reduce gaps in audio or video but add latency, which can degrade interactivity. Smaller buffers minimize delay but heighten the risk of clipping or glitching during congestion. The choice depends on the application’s tolerance for latency and packet loss, often supplemented by packet loss concealment methods.

Applications include VoIP endpoints, video conferencing clients, and streaming players. Jitter buffers are commonly part of RTP stack implementations and operate at the receiving side to stabilize media playout in the face of network variability.