IDK how sound processing went from 2020 to 2024.
For sure we do not want to process any sound on OTS (too much load on single CPU app). It must be off-loaded to some external server.
Skype did it - what you described - and sold for $8.5kkk in 2011 (P2P audio-video network).
Microsoft bought it and moved all client-side hosting/processing to server-side hosting/processing to hide IPs of users.
Add working noise reduction in Client and you got 1kkk$ company in 2024 ex. TeamSpeak.
If you got any working open source solution for this - even just for Windows - feel free to post. I would like to implement it in some OTClient/OTS.
TeamSpeak: 6 kb/s output/input to get all players clear voice (up to 100+ on single channel, server - VPS - merge voices into single sound stream; client removes all background sounds)
Open source (often designed for 2-5 connections): XX-XXXX kb/s output/input to get all players voice with all background sounds (noise)
TeamSpeak: full user/channel management by server
Open source: authentication? users? channels? not implemented
I must say, I never used TeamSpeak for any public OTS (only on 1-4 players test servers). With 100+ its price went to 30$+ per month and no OTS is able to pay it (why? they have free Discord!).
but I've discussed TS/other software with many OTSes and only option is TS (or Discord).