[gst-devel] Mixing silence suppressed RTP into a continuous stream

[Joakim Koskela]

Hi,

I'm having problems mixing an RTP stream with something else (testing with an 
mp3 file), as the RTP does silence suppression (sending packets only when 
noise is actually made). It seems that mixing is done only when *all* sound 
sources have data to contribute, leading to a chunky output.

What I would like to have is a continuous output stream into which sound is 
mixed from different sources as it becomes available. Seems like a really 
basic thing, but I somehow couldn't find anything using the search terms I 
thought relevant.

Any help is appreciated, even just tips on what to look at more closely (the 
adder or the sources) or keywords to search for. The pipeline i'm using is:

filesrc location=file.mp3 ! decodebin ! audioconvert  ! audioresample ! 
audio/x-raw-int,channels=1,rate=8000 ! adder name=mixer \
udpsrc port=%s caps="application/x-rtp,media=(string)audio,clock-
rate=(int)8000,encoding-name=(string)PCMA" ! gstrtpbin ! rtppcmadepay ! 
alawdec ! audioconvert ! audioresample ! audio/x-raw-int,channels=1,rate=8000 
! mixer. \
mixer. ! audioconvert  ! audioresample ! audio/x-raw-int,channels=1,rate=8000 
! alawenc ! rtppcmapay ! udpsink host=127.0.0.1 port=8000

br, j