Why is Jitter an Important WebRTC Metric for Contact Centers?

By Allie Mellen on January 21, 2019

Jitter is an important and widely used metric when diagnosing quality problems with real-time communications sessions. High jitter values contribute to poor audio quality, which can degrade customer experiences and prevent call center agents from communicating effectively. What is jitter, and why should WebRTC-enabled call centers keep track of it? Read on to find out.

If you’re interested in learning more about the importance of WebRTC in the call center, check out our white paper: Cloud-based Contact Centers: the WebRTC Story.

What is Jitter?

Jitter is the amount of variation in packet delay, which is why it is also frequently called delay variation. It is measured by calculating the average variation in packet arrival times at the receiving node occurring over a fixed interval. At callstats.io, the jitter calculation interval varies between five and ten seconds depending on endpoint configuration.  

In VoIP networks, packets are transmitted on the network continuously, typically every 20 milliseconds. They inevitably arrive at the receiver with differing delays, even when traveling the same path. Constant delay pacing simply cannot be guaranteed.

This differing delay can be a significant issue for real-time communications media streams, including call centers, because a quality reproduction of the sender’s voice requires data to arrive at the receiver at a constant rate

What Causes Jitter?

Two factors determine the amount of jitter in a media stream: packet loss and network congestion.

Congestion occurs when network routers receive more packets than they can transmit. When this happens, the router stores the excess traffic in a buffer and uses a queuing discipline to determine which of the packets waiting for transmission should go next. Thus, the amount of delay is a function of the amount of congestion, buffer size and queuing discipline.

Packet loss is caused by network errors, or excessive congestion. Examples of errors include unrecoverable link failures and connection handoffs that occur in wireless networks. Excessive congestion can cause a router’s buffer to overflow, resulting in packet loss.   

Congestion and packet loss cause packets to arrive at the receiver with varying amounts of delay, leaving the endpoint to deal with the problem of recreating a continuous listening experience.

Why is Jitter an Important WebRTC Metric for Call Centers?

Endpoints are designed to smooth out minor variations in arriving data by intentionally delaying the playout of incoming packets. Packets enter a jitter buffer that adds an imperceptible amount of delay to the listening experience, but allows the endpoint to create a continuous audio stream for the receiver. However, when jitter becomes excessive, the buffer can be emptied or it can overflow, causing choppy or scrambled audio experiences.

For an industry that relies on high-quality calls, this can be a huge problem. Poor audio quality can result in customer confusion, frustration, or even worse, churn. It also has the potential to extend time-to-resolution and lead agents to miss out on upselling opportunities.

How Can Contact Centers Compensate for Jitter?

Jitter problems can be addressed by relieving network congestion or increasing the size of jitter buffers, or both. Network managers need to determine which router(s) is experiencing congestion and increase its bandwidth or change the queuing discipline to expedite real time packets, as appropriate.

Endpoints may be equipped with one of two types of jitter buffers: static and dynamic. Static jitter buffers are implemented directly in the hardware of your system and are configured by your manufacturer, while dynamic jitter buffers are implemented in the software and configured by a network administrator. Dynamic jitter buffers can be increased to deliver a better listening experience for the receiver, albeit with a bit less interactivity in the conversation.

If your agents are experiencing poor audio quality that is impacting their conversations and customer experiences, jitter should be a metric you pay attention to.


Read A Case of High Latency

Tags: WebRTC, WebRTC Verticals, Contact Centers, WebRTC Use Cases, Amazon Connect