What is Echo?

Echo is your voice coming back to you, as if you were repeating yourself. During a normal two-person phone conversation, your voice is transmitted from your mouth to the ear of the person at the other end, and their voice is returned from their mouth to your ear. However, in any conversation, a certain amount of your own voice is also part of what you hear, whether you are talking face-to-face with someone who is sitting in your office, or talking to someone on the phone. This experience of hearing your own voice is not echo, but sidetone, and is a normal aspect of talking and listening.

Your own voice becomes echo when it comes to your ear with a significant delay from the time you spoke. Sidetone is scarcely noticeable when the delay between your speaking and hearing is less than 25 milliseconds. Within that time window, the human brain does not perceive the sound as echo.

Echo, then, is very much a function of delay. Once you can hear your own voice more than 25 ms later, the possibility of perceiving it as echo arises. Twenty-five to 150 ms is a typical delay range for international telephone calls, and this is why echo cancellation is necessary for such calls. Voice over IP calls don’t actually create additional echo, but they also have a delay budget in the range of 150 ms to preserve audio quality, so VoIP systems commonly employ echo cancellation as well.

Among the various settings and parameters that need tuning to preserve optimal call quality in a VoIP system, echo cancellation is one of the more complex, least understood factors that IP networking specialists must address when they encounter VoIP for the first time. To understand echo cancellation and the metrics associated with it, we need to look at some other aspects of echo that affect the implementation of echo cancellation on voice gateways.