Started by Zenith ·
"reticulum-party-line" is in NO WAY affiliated with the Partyline I have under active development https://github.com/RFnexus/partyline which does not use AI/ LLMs in any capacity outside of quick debug reference and future gettext localization.
It's the ciiiiircle of slop. Yet another disgrace.
Fwiw, I have an unreleased project that sought to combine LXST and RRC for group voice, but I never got around to polishing it enough to be worth releasing.
I haven't taken a close look at party line yet, but on the surface it seems to be a more thoughtful and careful approach to the same problem I was trying to solve (but I am coming from a discord world instead of mumble, so my UI looks quite a bit different).
Anyway, the one killer (to me) feature discord has that other group voice clients I've had to use is per-participant volume control. Turning up the one person who's too quiet or vice versa is so nice after a day of being stuck on MS Teams calls with one guy who turned his gain to 9000%. If party line has that already, awesome! And if it doesn't, do you think it's something that could be added?
Thanks Zenith! Genuinely glad someone else beat me to it lol, I am feeling stretched too thin as it is.
There are gain controls in the LXST primitives, but I'm pretty sure the partyline server downmixes all incoming streams to a single outgoing stream per connected client for bandwidth efficiency. It might be able to be done with some logic just detecting the speaker activity, and setting the gain on local playback sink, but I probably even shouldn't be trying to answer here, since I have no clue how Zenith wrote the playback architecture.
Mostly just wanted to say how nice it already it is. Haven't really talked to anyone yet, apart from myself and my family for playing around with it, but I have listened to a lot of traditional Russian music :)
Torlando wrote:
If party line has that already, awesome! And if it doesn't, do you think it's something that could be added?
Yes, there are client side per-user volume adjustments. I plan to (officially) release v1.0.0 sometime in the first week of October, but should be more than stable enough to play aruond with. I don't anticipate any protocol changes so it all should be g2g
Mark wrote:
There are gain controls in the LXST primitives, but I'm pretty sure the partyline server downmixes all incoming streams to a single outgoing stream per connected client for bandwidth efficiency.
Yeah, this is something I'd wrestled with in my own designing and experiments, and landed on a few config options to hopefully cover all the bases. Server hosts could configure whether they supported downmixing (maybe CPU constraints would keep them from wanting to) and clients could request the downmixed stream vs. separate streams per participant (trading client side mixing for better bandwidth), along with a quality request. The target use case for me is laptop/desktop group voice while playing computer games, so high speed links are mostly a given in my target use case, but I also wanted to support low bandwidth clients in case someone wants to participate via lora (or even a shitty hotel wifi connection).
Zenith wrote:
Yes, there are client side per-user volume adjustments. I plan to (officially) release v1.0.0 sometime in the first week of October, but should be more than stable enough to play aruond with. I don't anticipate any protocol changes so it all should be g2g
Awesome, I'll give it a try. I also love that you can dial in via an LXST client and that music channels are supported. I hadn't even considered those as possibilities :)
I'm waiting for it to be released so I can tell others to use it. Thanks a lot! I hope that for Version 1—or shortly after—you provide a one-click installer/executable for various operating systems. By the way, it would be great if the server setup process had a graphical interface... just a thought, in case you hadn't considered it.
