Usually the speaker is in a tiny rectangle in a corner, and the slides full screen. This is the "good" scenario, but even then you can't clearly see the speaker (because of dimensions, compression, or both).
Then there are progressively worse scenarios, like just streaming from a camera that was placed somewhere in front of the stage, recording a specific viewport. Or the speaker full screen and the slides in a tiny box in a corner.
If these things are separate in the stream/file, then the client software would be able to have options/layouts for what to display where, resize/move things, even split out in multiple windows (for multiple monitors).
Basically, a multi-layered video/presentation format.
PDF would make a nice layer to that, being able to change slides based on the timeline of the feed... or even change the slide manually, and it navigates the timeline to that point.
