[Asterisk-Users] De-Centralized / Distributed Conferencing App

[Adam Goryachev]

On Wed, 2004-10-27 at 06:50, Andrew Kohlsmith wrote:
> On October 26, 2004 04:15 pm, Richard Lyman wrote:
> > the issue i think is being discussed is when all participants can
> > talk.  if you were simply retransmitting client side muted then
> > you could fit alot of listeners amoung one or more servers to
> > handle the client load.  (meaning have the servers be the
> > conference member (1 channel) on the main server the input hits.
> >     basically .. 'streaming/multicasting'
> 
> I know that early on in the conference people could hear others -- there was a 
> lot of complaining about someone's cellphone which was drowning out the talk.

I think it would be helpful if there was a method for the user to mute
themselves from the conference, and toggle their mute status (maybe
pressing some DTMF digit to mute/unmute). Of course, even better would
be a method for each user to somehow "put up their hand", and eventually
the 'mic' would be passed to them. Even a first-in-first-served basis
might work.

ie, 300 people in conference, one person starts talking, after a
configurable time-limit, or when the person presses a button, they 'mic'
is removed and handed to the next person who indicated they wanted to
speak. A record is kept of the order that the mic should take. A user
can also cancel their request before their turn, or release their turn.

This would allow for a 'self-moderated' conference...

Of course, even the first option, where you can 'push-to-talk' would
drastically improve things (filtering background noise from 300 people).
I suppose if you wanted to get fancy, you could do speech detection on
the 300 channels to auto-detect their 'push-to-talk' status, but this
would (I think) be quite difficult. (This is different to simple silence
detection I would think, but that might be a medium-term solution).

Regards,
Adam