Hello all!
I don't know why, but it seems that am I the only one who suffers from
echo problem? Probably not. But several implementations or * are
currently waiting, because of the echo problem.
Here is the list of measures, I tried out:
1. I have used GRANDSTREAM BUDGETONE 101 phones. They are declaring,
they have implement G.168 cancellation measures, I at first I decided to
go with them. I have tested them and it seems they quite effectively
nullify echo from most (analog) systems I have tested. They are quite
successful with analog PBXes, being connected over several ISDN BRI
connections. But I had great troubles with ISDN DECT (cordless) phones.
SIEMENS is having several models and as far as I can tell you, echo
stays in many situations I have experienced. The problem is that here in
Slovenia is, such cordless phones were sold *EN MASS*, so still in 3 out
of 10 cases, the echo problem persists. Besides, I have to use just BT
101 phones and I am not quite flexible here. I planned to use many SNOM
phones. I'm not sure, if I am experiencing the acustic echo with SIEMENS
ISDN BRI DECT phones or not. I seems to me it is the electrical echo,
but I cannot imagine, where this echo comes from, since there is no wire.
2. I searched for some kind of boards or whatever, for me to put them
in, in order to nullify echo. No luck. Boards were either quite
expensive (and not ment for * at all), or I was quite doubtfull, that
they accually do the work. I'm pretty amazed, that no such solution
exists. I'm aware, the problem of echo must be solved as much as close
to the source channel (ISDN BRI in this case). But this way, every
driver have to have their own echo cancelling measures supported and I
think this is not the most optimal way. But I have not found the board
anyway.
3. I have used the bristuff. They are stating, they are have echo
cancellation implemented. I already tried bristuff before, and I had
compilation problems. Interestingly enough, when using * v1.2.1 and
bristuff 0.30.0, no compilation errors gotten. I have implemented the
usual staff - extensions, sip.conf etc etc.
4. And yes, I have looking to use the echo cancellation in vocodec. iLBC
guys are declaring, they have counter measures implemented (i.e. echo
cancellation). Unfortunately, when implementing PBX solutions, I usually
use the G.711 mu codec, since it gives the best (and problem less)
results in local LANs. Besides, BT and SNOM phones do not support iLBC.
So, I had no luck with codec echo cancellation techniques.
As far as I can tell you, bristuff cancells echo effectively. No echo
whatsoever. I have left all echo cancellation parameters intact - as
they have been declared from the beginning, not changing the number of
taps, etc. etc. 128 taps nullify echo down to none. This works.I have
not tested *international* TDM calls - where I presume a bit more of the
delay will enter the game.
I know that this is the vISDN list, but... I will stay on vISDN list
and will be ready to test the echo canceller, implemented in vISDN. But
that echo problem bothered me quite a lot and finally I got the
solution. Pheeewww....
I will be testing echo cancellation even more and will report to this
list if important matters will occur.
=b
Newer Digium hardware has hardware echo cancellation, however, most boards
still rely on the software EC in zaptel drivers. I'm taking that canceller
and putting it in vISDN.
Its performance is not exceptional but would be way better than going without
an EC :)
_______________________________________________
Visdn-hackers mailing list
[email protected]
https://mailman.uli.it/mailman/listinfo/visdn-hackers
