It's Covid days and I am sure all of us who play Pen and Paper RPG's (watup, nerds) have found the experience incredibly lacking.
The way we communicate via Discord etc, has to be incredibly different from IRL conversations. Its frustrating when the core element of RPG's is the conversations, the chat, the small talk, the adlib and the silly jokes.
The way we talk IRL is so different because we can discern the different sources, we can listen more or less to different people, we can interrupt and add things. Conversations via Discord is more like listening to a speech, and then replying. IRL gaming and the conversations that crop up are more like actual human chatting - taking a joke, building on it, having it taken from you etc etc. (the way me and my friend talk is so natural, we know each other well enough to be able to discern the relevant from the irrelevant - the bits we can tag on to, and the bits we need to leave alone)

What I was thinking was to see HOW we talk in gaming, and how that could be mimicked SOMEWHAT in Discord etc.

My idea was to create a set of icons/low quality videos arranged around a table placement, that you can then focus on. Like a mouse controlled object that indicates what part of the table you focus on and how much. Just like a human would by turning her head this way or that to focus on one person talking in a group, or leaning in towards that source to indicate how much she listens to that unique source. By having a physical placement you can focus on an edge of "the table" and then lean in towards one source - and lean out to listen to all.

All other listeners can see your focus, the way you turn towards a source and be able to change their communication to fit.

So imaging having your "icon"/video at the bottom, the table sorted in a half circle shape above and your "focus" in the middle. If you pull it to its "lagrange point" (a snapping midpoint so its easy to find) you are listening to all. By dragging it towards one end you are focusing more on that person and the people around it, as a circle. You can see others focus, by coloured lines focusing more or less on an end or another.

My idea is to abuse the already available 3D audio effects existing and use that to put your focus towards one end or another, muting and muffling audio as your focus move across the table to be able to somewhat mimic the way we as human listen.

The social order of an RPG session, with the DM being the natural focus at times means that that focus can happen naturally either through focus, or simply silence. With it you can find the focus of others as an indicator of whether you are committing a faux pas or not, just like in real life when people pointedly may look directly towards another source.