I have a point-to-point sender receiver pair coded to send audio only between two endpoints, as in an intercom. My encoding is not a well know Wowza standard; it is
live microphone -> PCMA encode -> RTP payload -> UDP transport sender
which is decoded on the other end as
UDP reveiver -> RTP depayload ->PCMA decode -> speakers.
The code is a Gstreamer pipeline on both ends.
Now I want to pass my "intercom" through a Wowza server and it looks like I'm going to have to abandon PCMA to accomplish this.
The Wowza Live stream's input side will accept mpeg1 so that the gstreamer sender pipeline will become
microphone (autoaudiosrc) -> Mpeg Audio Encoder (mad) -> mpeg ts muxer (mpegxmux) -> UDP sender (udpsink).
Doing this I get back an mpegts stream that I can easily decode with ffplay on Windows and Linux or a gstreamer pipeline
rtmpsrc ! mpegtsdemux ! mad ! autoaudiosink
on Linux, but not on Windows where the source and/or demuxand/or decoder pipeline elements are broken.
And that's my problem. I prefer a single solution that works on both OSes.
Before I go further down any non-gstreamer path, I wondered if there is any way to get the transcoder (my best guess at the right add-on to use) to map my input stream back out in the same format it received it. Or at east send it out as RTP not RTMP? I'd prefer PCMA, very low bitrates and LGPL.
Or would the Push-Publish add-on be an easier approach?
Can someone get me pointed to a reasonable starting place or share an example of doing such a thing?