[Asterisk-Users] EC needed in all-digital situation?
Warren
warren-lists at icruise.com
Thu Jun 15 12:35:56 MST 2006
Andrew Kohlsmith wrote:
>On Thursday 15 June 2006 10:22, Warren wrote:
>
>
>>I was just told that for my forthcoming system I will be getting a data
>>T-1 instead of a voice T-1. Given that all of the handsets will be voip
>>phones, no analog at all, do I need echo cancellation? I looked at the
>>voip-info wiki and it seems to me that the answer should be "no" but I
>>would like to confirm that.
>>
>>
>
>All-digital setups do not *generate* echo since there is no hybrid circuit to
>reflect energy. However all-digital systems can still have echo if the
>far-side is a 2-wire system and the latency is sufficiently high. (Anything
>with a PCI bus can bring this latency up easily.)
>
>Also, if you have an ueber-cheap phone which acoustically couples the speaker
>in the handset with the mic you can introduce echo on your all-digital system
>that way. The same goes for cheap speakerphones and/or acoustically "hard"
>rooms.
>
>-A.
>_______________________________________________
>--Bandwidth and Colocation provided by Easynews.com --
>
>Asterisk-Users mailing list
>To UNSUBSCRIBE or update options visit:
> http://lists.digium.com/mailman/listinfo/asterisk-users
>
>
So the next question becomes... Is hardware EC necessary or can *
handle it in software? I am looking at some pretty beefy hardware for
my platform, a Dell PE2850 with dual Xeon 3Ghz processors and plenty of
RAM to spare. I am likely starting with a single T-1 PRI line and
expending to 2 within a year, but that would be the max. I am not sure
of how many simultaneous calls that would be, but let's assume I want to
use a b-channel's worth of bandwidth for each, so 23 max to begin with.
Can * do software echo cancellation for 23 (and eventually 46) calls on
a dual Xeon 3.0Ghz system?
W
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20060615/5e9e596b/attachment.htm
More information about the asterisk-users
mailing list