screencasting: Minimise pixel format conversions while streaming

Instead of best-guessing, at BGR (which in retrospect was a bad guess), offer whatever resembles most the internal representation. This way the frame gets to be least treated as it goes into the client.

It includes !3594 (merged), so please ignore in the context of this MR. It's getting merged as soon as the BSD runners decide to build our stuff.

Draft because we need to make sure the format we choose is compatible with dmabuf too