[Asterisk-Users] Why echo occurs

[Steve Underwood]

Hi Eric,

Before VoIP and digital cellular systems, audio samples passed down the 
communication path with little more delay than the pseed of light. A 
couple of samples delay might be incurred in switching equipment, but 
most operators demand a 3 sample maximum hardware delay down the line. 
Now audio samples are bundled (typically in 20-30ms chunks) for 
packetising, and to go through low bit rate codecs. Then the bundles get 
queued in IP routers, creating further latency.

VoIP uses massive amounts of complexity in a struggle to offset the 
effects of these delays, and approach the quality older telephony 
achieved with simple equipment. In the end the simple equipment will 
always beat it. That said, in another 10 years there will probably be no 
traditional PSTN. Strange, huh?

Regards,
Steve


Eric Bishop wrote:

>OK I understand that the $5 handset may indeed have an echo but that
>it occurs so fast that it is not preceived as an echo. I pose the
>following questions:
>
>1. Is the echo (regardless of it's speed) a side effect of long
>distance communications or is it there by design for some technical
>purpose?
>
>2. Is only a problem in 2-wire technologies (ie analog and BRI ISDN lines)?
>
>3. Where exactly is the slowdown occuring? For example take my Supira
>3000 as a case in point. It takes no longer for the PSTN signal to
>reach the Sipura's FXO port than it does my $5 handset. Going from the
>other end it takes no longer for the SIP signal to reach to the
>Sipura's ethernet port than it does any other IP phone. So logically
>the slowdown is happening as Sipura converts the PSTN signal to SIP
>and so forth. Is it just that the Sipura/TDM400 etc. have a too slow
>conversion CPU. Would a faster digital to analogue audio converter
>"fix" the the problem?
>  
>