asterisk users - Feb 2006 - Meetme echo cancellation

Hi there

I am using IAX2 softphones dialing into a meetme conference. In my softphone
I was forcing uses to click on a button when they wanted to speak, enabling
their microphone and disabling their speakers. This way when a user was
speaking they did not hear their voice half a second later (because meetme
mixes the voice and sends to everyone in the conference).

Now because of requirements there is a need for users not to have to click a
button when speaking (and have their microphones and speakers enabled at all
times) - much like Skype. How would I prevent a user hearing their own voice
half a second later? Using some kind of echo cancellation? I am not sure
that this is defined as echo though.

Does anyone have any ideas?

Many thanks

Steven
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.digium.com/pipermail/asterisk-users/attachments/20060209/d91ca8c7/attachment.htm

Steven Langley wrote:
> I am using IAX2 softphones dialing into a meetme conference. In my
softphone
> I was forcing uses to click on a button when they wanted to speak, enabling
> their microphone and disabling their speakers. This way when a user was
> speaking they did not hear their voice half a second later (because meetme
> mixes the voice and sends to everyone in the conference).
This is wrong. app_meetme does not send the speaker's own voice back to
them, because the mixing in Zaptel removes it for the speaker's channel
in the conference.

Try it :-)

In article <20060209155523.90FC7370E@ctb-mesg9.saix.net>,
Steven Langley <steven@intellinc.co.za> wrote:
> 
> I am using IAX2 softphones dialing into a meetme conference. In my
> softphone I was forcing uses to click on a button when they wanted to
> speak, enabling their microphone and disabling their speakers. This way
> when a user was speaking they did not hear their voice half a second
> later (because meetme mixes the voice and sends to everyone in the
> conference).
> 
> Now because of requirements there is a need for users not to have to
> click a button when speaking (and have their microphones and speakers
> enabled at all times) - much like Skype. How would I prevent a user
> hearing their own voice half a second later? Using some kind of echo
> cancellation? I am not sure that this is defined as echo though.
> 
> Does anyone have any ideas?
It's a hard problem, if your users are using speakers instead of headsets
or normal phone handsets.

On traditional speakerphones, it is the phone itself that has echo
cancellation built-in, so that any speaker output that gets picked up by
the microphone gets cancelled out in the signal that is sent along the
phone line.

So really, it is your soft phone that should be replicating the same
functionality, and making sure it does not feed back audio from the
speakers.

I think It would be much more difficult to try to do it at the asterisk
end.

Cheers
Tony
-- 
Tony Mountifield
Work: tony@softins.co.uk - http://www.softins.co.uk
Play: tony@mountifield.org - http://tony.mountifield.org