Now that all platforms use gstreamer 1.x, there is no reason not to use opus if both ends support it. I have used it before and the code tunes it for low latency.
In theory, it will help us with av-sync (#835) because of its much lower latency:
bitrate comparison image from wikipedia
quality comparison from wikipedia
Added opus in r11597 builds on my old build machine and the new one.
Rebuilding gst-plugins-bad-1.0 should now include this but I will also make it a dependency so it gets rebuilt.
(@smo: I assume this is done, re-assigning to afarr)
afarr: testing: latency for av-sync, buffer levels, cpu usage, bandwidth usage
I believe that the packaging bits are in place as of r11761.
@smo: please ensure the modulesets have an opus dependency somewhere before handing over to afarr.
Raising as blocker for the release.
Testing right quick with a 0.17.0 r12401 osx client against a 0.17.0 r12453 fedora 23 server, both seem to indicate
opus+gdp first on speaker and microphone codec lists (making them the defaults, if I don't miss my guess)... and sound works well.
I don't see anything in the session-info to make it any easier than before to determine which codec is actually in use, other than assuming that the first listed is the one in use.
Unless I'm missing a detail, I think this is ready to be closed.
I don't see anything in the session-info to make it any easier than before to determine which codec is actually in use
You can see which codec is used in the command output or via "xpra info".
I'll try to stick on session info somewhere.
I must be the first one to test opus with Ubuntu Xenial, immediately found a bug, fixed in r12468.
For OSX decoding support, see also: ticket:1204#comment:1
this ticket has been moved to: https://github.com/Xpra-org/xpra/issues/1074