Using GStreamer to Convert WAV -> RTP Results in Low-Quality RTP Stream

Ian Davidson id012c3076 at
Tue May 16 14:33:34 UTC 2017

I see that you are creating a stereo output (channels=2), but it is 
being played as a mono (channels=1), so each sample from the original is 
being treated as 2 samples.  This means that it will take twice as long 
to play as it did to record.

I don't know why.


On 12/05/2017 18:50, Josh Dickson wrote:
> Hi Jan,
> Thank you, that is definitely what I need. I have gotten that pipeline 
> working successfully, but now when I play it (via ffplay), it sounds 
> comically slow/distorted.
> I am now using the pipeline:
> gst-launch-1.0 -m filesrc location=myfile.wav ! wavparse ! 
> audioconvert ! rtpL16pay ! udpsink host= port=12008
> I used the -v option to produce what I thought was a correct SDP file:
> v=0
> o=root IN IP4
> c=IN IP4
> s=My Name
> m=audio 12008 RTP/AVP 96
> a=rtpmap:96 L16/44100
> a=fmtp:96 media=audio; clock-rate=44100; encoding-name=L16; channels=2;
> I am playing the sound with:
> ffplay -i stream.sdp -protocol_whitelist file,udp,rtp
> Ffplay does open, and the sound resembles the original song, but it is 
> very slowed down/distorted.
> Ffplay sees:
> bitrate: 705 kb/s
>     Stream #0:0: Audio: pcm_s16be, 44100 Hz, 1 channels, s16, 705 kb/s
> (not sure if that will help)
> I have been trying to research what is wrong here but I am not sure 
> what part of this I’ve messed up. Any help would be much appreciated. 
> Thank you!
> Josh

More information about the gstreamer-devel mailing list