Jitter Buffer Overview

Voice over IP (VoIP) carries a call as a stream of small data chunks called packets. The network rarely delivers those packets on a perfectly even schedule, and uneven delivery is heard as choppy or garbled audio. This page explains what a jitter buffer does, the two buffer types Cloud Voice provides, and when turning one on helps.

Understanding jitter

Jitter is the variation in timing between when voice packets leave the sender and when they reach the other end. When the network is congested, packets can arrive bunched together, spread apart, in the wrong order, or even at the same time as one another. Any of these disrupts the audio and lowers call quality.

A jitter buffer counteracts this. It holds each incoming packet for a very short time, then releases the packets in line with their expected timing, so the audio plays back as a smooth, evenly spaced stream.

Note

A jitter buffer treats the symptom, not the cause. It smooths out how packets reach the audio stream, but it does not fix the underlying network congestion, bandwidth shortage, or bad connection. If audio problems persist, investigate the network itself.

Buffer types

Cloud Voice offers two jitter buffer implementations:

• Fixed jitter buffer: Uses a set buffer size. Every packet that leaves the buffer carries the same, constant amount of delay.
• Adaptive jitter buffer: Changes its size in response to the network’s current delay, so the delay applied to outgoing packets varies over time.

Tip

An adaptive buffer suits networks whose conditions change from moment to moment, because it grows and shrinks to match. A fixed buffer suits networks whose delay is steady and predictable. When you are unsure which to pick, adaptive is the safer default.

Caution

Both buffer types add delay (latency) to the call, since packets are held before they play. A small amount is normal and expected. If the buffer is larger than the network actually needs, the extra latency can make the conversation feel laggy, with each side talking over the other. Enable jitter buffering only when you are troubleshooting audio problems, and do not set it larger than required.

When to enable a jitter buffer

Turn on a jitter buffer when the network is causing audio problems such as lost packets or packets that arrive in the wrong sequence. The buffer handles each case as follows:

• Packet loss: When some packets go missing, the buffer inserts a replacement for the lost frame (one small unit of audio) and forwards the audio as a steady, evenly spaced stream.
• Out-of-order packets: When packets arrive in the wrong sequence, the buffer reorders them correctly before passing them along in the order the listener expects.

For the steps to set this up, see Configure Jitter Buffer.