[Asterisk-Users] X100p echo guide

[Rich Adamson]

> Nice response.
> 
> I just wanted to let you know that we do provide a tool to debug echo. 
> 
> We send a unit impulse and record the Finite Impulse Response (FIR) so it
> can be plotted and analyzed. The code that does this is the release at
> ftp.sangoma.com/linux/custom/2.3.4. Instructions on using it are in the wiki
> in http://sangoma.editme.com/wanpipe-linux-asterisk-debugging.

Excellent, I'll give it a try. I'd love to see some tools that would
help those that are not necessarily familiar with the issues.

> Although the code is wanpipe, all the interaction is at the zaptel level, so
> I am pretty sure it will work on Digium or other cards as well.
> 
> Just being able to see what the echo looks like on a troublesome line gives
> quite a lot of info. You can see if the echo is delayed, or markedly
> non-linear.
> 
> The FIR plots look like the ones in
> http://www.linuxjournal.com/article/8424. 
> 
> BTW, a small correction: The "echocancel=128" defines how large of an echo
> canceller to use in taps (1/8 milliseconds) not milliseconds. echocancel=128
> gives you a 128 tap (16ms) tail. Telco echo cancellers give a full 128ms =
> 1024 taps.

Ops, that's what I get for slamming a few beers while doing email. ;)

I'd really love to contribute to the cause by documenting many of the
pstn interface issues, but adequate tools to help the less knowledgable
assess the environment tends to be a shortcoming.

Rich