Jitter Explained

Jitter: Inconsistency, Not Just Slowness

Latency (covered in the previous note) measures how long a delay is. Jitter measures something different: how much that delay varies from one packet to the next, even between packets sent close together.

Packet 1: arrives after 20ms
Packet 2: arrives after 45ms
Packet 3: arrives after 22ms
Packet 4: arrives after 60ms
     (highly inconsistent delay = high jitter, even though average latency might seem fine)

Why Inconsistency Can Be Worse Than Consistently High Latency

A steady 100ms delay is often easier to work with than a delay that unpredictably swings between 20ms and 150ms - many real-time applications can adapt to a known, stable delay far more gracefully than to unpredictable, constantly shifting delay.

How Applications Compensate: Jitter Buffers

Voice and video call applications commonly use a small "jitter buffer" - deliberately holding incoming data for a brief moment before playing it, smoothing out arrival-time variation into a steady, consistent playback stream, at the cost of adding a small amount of extra delay.

What Typically Causes Jitter

• Variable network congestion - traffic levels fluctuating from moment to moment along the path.
• Packets taking different routes - especially relevant on networks where routing can shift dynamically.
• Inconsistent queuing delay - packets sometimes waiting briefly in router buffers, sometimes not.

Why Jitter Is Specifically Critical for Voice/Video

Unlike a file download (where jitter has essentially no practical impact, since all the data eventually arrives and gets reassembled anyway), a live voice call has no opportunity to "wait and catch up" - which is exactly why VoIP quality, covered later in this project, depends heavily on managing jitter specifically, not just average latency.