docker-jitsi-meet
docker-jitsi-meet copied to clipboard
Add ability to Jigasi's sip-communicator.properties to avoid one-way communication with (cell) phones
The Problem: When certain SIP dial-in accounts are used, a participant using a phone (with me, it's cellphones only) can speak, but hears nothing. A fix has been found, e.g. see in this Forum entry. The entry probably solves other issues, too.
Up until now, I have been helping myself with patching the sources manually before creating the docker images.
To streamline seployment, can you please make this setting configurable through .env / docker-compose.yml? Thanks!
org.jitsi.impl.neomedia.transform.csrc.CsrcTransformEngine.DISCARD_CONTRIBUTING_SOURCES=true
@damencho thoughts?
This has nothing to do with one way audio ... If I remember correctly. This skips adding an extra rtp header which some servers were not reading correctly ... Actually on second thought, yes this can be one way audio when those failed to read the rtp send from jigasi.
Hum I thought this is on by default.
Reports for freeswitch were noisy sound not completely no audio ...
https://community.jitsi.org/t/continuous-noisy-sound-at-sip-client-jitsi-meet-conference-to-sip-client/31040?u=damencho https://community.jitsi.org/t/solved-jigasi-crackling-sound/28965/2?u=damencho https://community.jitsi.org/t/buzzering-sound-on-landline/24004?u=damencho
You can turn it in by default or I can add it on by default in jigasi
@damencho , thanks for looking into this.
Not being a specialist, I don't know what it does technically, but I have ample evidence (3 servers currently running) that setting the option to true fixes the one-way audio issue to cellphones (landline works fine, for whatever reasons). And therefore default must be "false", because every install I did (most recent: stable-5870) started with one-way audio, and was fixed by me.
Since other issues seem to be fixed with this, too, it might be a good idea to set the default to true, and provide an option to disable it.
Actually, we cannot turn off that by default as it will stop sending the audio levels when those exist (rfc6464). So it makes sense to add it as a property for the docker image and in case someone is not using audio levels and have a buggy sip service to be able to turn it off.