Pete,

   Do you have a hard requirement to handle this programatically?  If not,
I would use a tool such as Adobe Premiere Pro to just create the movie that
you need to create and export it in AVI (or whatever) format you need.
Then, you can just display the video in your app.

   If you do have this hard requirement, then I would assume you can create
a video from a series of images and some audio, but you will have to code
the rendering at the appropriate time, etc, etc.

   You should be able to get PCM data from direct X.  This would most
likely involve creating a filter which is then inserted into the filter
graph in the rendering of the audio.

   XNA is used for creating games, and is not specifically for multimedia
like direct x is.  You are better off going the DX route.  XNA does use
DirectX underneath the covers (I am almost sure of this), though, but it is
really an abstraction meant to make game development easier.

   I am sure on some level XNA supports UI elements, but to what extent, I
don't know.