[Freeswitch-users] One to many w/o mod_conference

Michael Jerris mike at jerris.com
Wed Oct 11 20:32:06 UTC 2017


its not copying rtp packets, its copying decoded audio.

> On Oct 11, 2017, at 4:16 PM, Tom Hartnett <hartnett.tom at gmail.com> wrote:
> 
> Thanks for the clarification Michael. I was working off this old post
> http://lists.freeswitch.org/pipermail/freeswitch-dev/2009-June/002414.html <http://lists.freeswitch.org/pipermail/freeswitch-dev/2009-June/002414.html>
> Which I guess is wrong or outdated.
> 
> On Wed, Oct 11, 2017 at 3:46 PM, Michael Jerris <mike at jerris.com <mailto:mike at jerris.com>> wrote:
> eavesdrop is going to transcode.  its copying the raw already decoded audio from one channel to the other then re-encoding it.
> 
> 
>> On Oct 11, 2017, at 3:36 PM, Tom Hartnett <hartnett.tom at gmail.com <mailto:hartnett.tom at gmail.com>> wrote:
>> 
>> Sorry to beat a dead horse, but so you mind if I ask why? The info I could discern (which may or may not be accurate) was that eavesdropping uses a simple RTP copy mechanism, while conferencing would involve converting all audio back and forth to PCM for each user (as well as jitter buffers and other overhead for each participant). Is there something I'm missing?
>> 
>> On Wed, Oct 11, 2017 at 2:56 PM, Michael Jerris <mike at jerris.com <mailto:mike at jerris.com>> wrote:
>> mod_conference would be at least just as efficient as that, probably more.  The feature you are looking for doesn’t exist today.  
>> 
>>> On Oct 11, 2017, at 2:42 PM, Tom Hartnett <hartnett.tom at gmail.com <mailto:hartnett.tom at gmail.com>> wrote:
>>> 
>>> Thanks for your comment Michael,
>>> Would it be reasonable to have a bunch of callers (listeners) spy or eavesdrop on a single call to a dummy endpoint (sender)?
>>> 
>>> On Wed, Oct 11, 2017 at 2:39 PM, Michael Jerris <mike at jerris.com <mailto:mike at jerris.com>> wrote:
>>> we currently do not have a media distribution method to other endpoint modules other than mod_conference.  You can use one of the stream recording methods to distribute to cdn as long as the delay isn’t an issue.
>>> 
>>> > On Oct 11, 2017, at 12:38 PM, Tom Hartnett <hartnett.tom at gmail.com <mailto:hartnett.tom at gmail.com>> wrote:
>>> >
>>> > Greetings all,
>>> > I have an application where I could use some expert advice. I'd like to set up an audio broadcast bridge, where the audio from one call would be distributed to multiple listeners. Listeners have no send audio. Think a corporate conference call with only one host. All listener calls are incoming.
>>> > I could of course set this up with mod_conference and just mute all the listeners. But this requires a bit of overhead to convert everything to PCM then back to the listener's codecs individually. I'm working on a resource-strained embedded system, so I'd like to find a way to simply bridge the speaker audio to multiple endpoints simultaneously. I have very tight control over the codecs used, and can make sure they are all the same.  Sort of like live MOH (I'd be using mod_portaudio for the sending endpoint in fact). Delay must be very low, so converting to a shoutcast stream and using the MOH function isn't an option either.
>>> > Can anyone offer any advice? Maybe FS is just not suited to this?
>>> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freeswitch.org/pipermail/freeswitch-users/attachments/20171011/9d5aeb1e/attachment.html>


More information about the FreeSWITCH-users mailing list