Raspberry PI 4 Selective Forwarding Unit - 3 way full duplex audio conference via SIMPLE-RTSP-SERVER скачать в хорошем качестве

Raspberry PI 4 Selective Forwarding Unit - 3 way full duplex audio conference via SIMPLE-RTSP-SERVER

*no pi sound card is needed to setup the PI as a SELECTIVE FORWARDING UNIT(SFU) in this example demo... 1st the RTSP SIMPLE SERVER is setup on the PI https://github.com/aler9/rtsp-simple-... 2nd a Gstreamer MULTI UDP incoming forwarding script is written to decode/RE-encode and then forward every UDP incoming CONNECTION over to the pi's RTSP-SIMPLE-SERVER input.... NOTE: each UDP(person) will have their own unique channel pi@raspberrypi:~ $ gst-launch-1.0 rtspclientsink name=s location=rtsp://localhost:8554/mystream udpsrc port=2222 ! "application/x-rtp, media=(string)audio, clock-rate=(int)48000, encoding-name=(string)OPUS, sprop-maxcapturerate=(string)48000, sprop-stereo=(string)0, payload=(int)96, encoding-params=(string)2" ! rtpopusdepay ! opusdec ! audioconvert ! opusenc bitrate=256000 ! s.sink_0 udpsrc port=3333 ! "application/x-rtp, media=(string)audio, clock-rate=(int)48000, encoding-name=(string)OPUS, sprop-maxcapturerate=(string)48000, sprop-stereo=(string)0, payload=(int)96, encoding-params=(string)2" ! rtpopusdepay ! opusdec ! audioconvert ! opusenc bitrate=256000 ! s.sink_1 udpsrc port=4444 ! "application/x-rtp, media=(string)audio, clock-rate=(int)48000, encoding-name=(string)OPUS, sprop-maxcapturerate=(string)48000, sprop-stereo=(string)0, payload=(int)96, encoding-params=(string)2" ! rtpopusdepay ! opusdec ! audioconvert ! opusenc bitrate=256000 ! s.sink_2 to send to the PI's UDP rtspclientsink forwarding script... each person uses different ports with the same basic udp send script like this: gst-launch-1.0 -v autoaudiosrc ! "audio/x-raw, channels=1" ! audioconvert ! opusenc bitrate=128000 ! rtpopuspay ! queue ! udpsink host=(ip address of the pi) port=4444 to receive all connections from the Pie'Z rtsp-simple-server, a script like this is used: gst-launch-1.0 rtspsrc latency=0 location=rtsp://(ip address of the pi):8554/mystream name=src src. ! rtpopusdepay ! decodebin ! audioconvert ! audioresample ! jackaudiosink sync=false async=false src. ! rtpopusdepay ! decodebin ! audioconvert ! audioresample ! jackaudiosink sync=false async=false src. ! decodebin ! audioconvert ! audioresample ! jackaudiosink buffer-time=11000 sync=false async=false src. ! rtpopusdepay ! decodebin ! audioconvert ! audioresample ! jackaudiosink sync=false async=false NOTE: you may need to add an RTPJITTERBUFFER element before the rtpopusdepay elements, if there are too many audio glitches/dropouts etc...