rtp-forwared audio to gst-streamer

[Дмитрий Тютерев]

Please help to unedrstand rtp format
I used videoroom plugin and successfull use such piplelines
gst-launch-1.0 -q udpsrc port=UdpPort ! "application/x-rtp,media=(string)audio, clock-rate=(int)48000, payload=(int)111,encoding-name=(string)X-GST-OPUS-DRAFT-SPITTKA-00" ! rtpjitterbuffer latency=10 drop-on-latency=true do-lost=true mode=0  ! rtpopusdepay ! opusparse ! opusdec plc=true ! audioresample ! audioconvert noise-shaping=4 ! alsasink device=plughw:0,0

Later when i try to use audiobriedge with sampling 48000 it's work too
But if I try :
1) Send rate to 24000 and chanche caps to "application/x-rtp,media=(string)audio, clock-rate=(int)24000, payload=(int)111,encoding-name=(string)X-GST-OPUS-DRAFT-SPITTKA-00"
2) Use opusparse pipeline
3) Use gst-discoverer-1.0 utility

I have diff trouble

Can you get some information about stream witch i have after rtp_forwarding

And THANX for janus_gateway !

[Lorenzo Miniero]

Opus only does 48kHz, different sampling rates are handled transparently and inband. You have to downsample it yourself with an audioresample if you want a different one in your pipeline.

L.

[Дмитрий Тютерев]

What is 'sample_rate' in audiobridge plugin for ?

понедельник, 3 октября 2016 г., 15:24:43 UTC+3 пользователь Дмитрий Тютерев написал:

[Lorenzo Miniero]

Internal and actual sampling rate that is used for Opus. Opus frames are indeed capped at that rate, but from an RTP perspective it still looks and is transmitted like 48kHz though.

L.