Mixing audios from N-to-N conversation

[Yokemuratakeshi]

Hi, all

This is related to  my previous question, in which 

I asked how to build a few-to-many live streaming system, 

and I understood that it's possible by building N-to-N conference and then

on top of that streaming each N streams to audiences separately.

This is totally fine for relatively rich viewing environment,

but as I'm also targeting mobile environment as well,

and in those environment receiving multiple streams at the same time may be too heavy 

because of both computing power and signal bandwidth.

So what I'm thinking of right now is to mix the steam before distributing to the audiences.

Fortunately my system is audio only, there will not be many options in how the streams should be mixed.

Of course there will be more lags on the audience side as another processing process is added to the pipeline,

but I believe this is going to be better than consuming too much bandwidth on audience side.

It seems like this cannot be done only with Ant's functionality,

but I guess it's still possible by combining external applications, services or such.

Does anybody know a good way to do that?

Any info is appreciated.

Thanks!