Always deal with MediaItem at the LayoutMedia layer

[hughns]

That way we only convert to the view model when we do LayoutMedia => Layout

[toger5]

I see how this simplifies things. If I understand correctly one can phrase the PR as shifting the transition from Media to MediaViewModel into a lower level.
Passing the full MediaItem instead of just the ViewModel increases the access however. I think the reason for creating it like this is to have as limited access to the MediaItem as possible in the CallViewModel.

The vm should represents this "smallest set of data".

What is the win we get by passing the whole MediaItem?
If there is too much redundancy I think we might need to evaluate if MediaItem is already a View Model. And we should copy all UserMediaViewModel code into UserMedia.

Is there any data from the MediaItem that we need in the Layout classes?