Network Bandwidth Usage and QoS

I have questions about bandwidth usage behaviors:

  1. When a user shares a still image and does not move cursor on it, we expect to see a decrease in both outgoing network traffic of that user and incoming of others. Is there any way to achieve this?
  2. Is Audio bandwidth usage directly proportional to the number of attendees with mics on even they are not speaking? Is there any way to decrease or limit this traffic? I know Jitsi uses SFU, but is there a way to use an MCU like processing for audio channels?
  3. Does tweaking with Codecs help us to decrease bandwidth usage? If so, what are your recommendations?
  4. Is there any way to guarantee QoS on audio in order to avoid issues like robotic sound etc?
  5. What should we do to reduce lipsync issues during video calls?

Thanks in advance
Argon

For item 2 -
This seems not to be possible.Please refer to:
https://community.jitsi.org/t/server-side-background-noise-filter-using-nvidias-rtx-voice-filter/44262.

Maybe it is possible to implement some logic at the client side, like closing audio track when no voice is detected for a period of time; and switching it on when conversation starts at that client side. There