I am looking for developing a multi-touch rhythm game in C #. So far, I have decided several requirements for the game, which must be present absolutely:
1) Unicode text display, without characters that are previously known (i.e. displayed lines are not known at compile time, but entered by the user).
2) Support for displaying all common image formats (PNG, JPG, BMP, GIF) and playing various audio (MP3, WAV, OGG, FLAC) and video formats.
3) Full support for multi-touch (I would like to support as many simultaneous inputs as the user's hardware allows).
I played with XNA, but I found it difficult to draw arbitrary Unicode text using SpriteFonts (he hates it when you try to download the entire Unified CJK set). Also, I did not find an easy way to load certain formats, such as Ogg Vorbis sound in XNA.
So, with that in mind, I began to learn how to use DirectX 10 through SlimDX. However, I am very lost and not sure how to start with it, and if the opportunities that I need are even present in it. To top it off, SlimDX documentation seems to be sorely lacking, especially when I'm new to game development. Which of the two frameworks is best for my purposes?
EDIT 1 . Any links to the initial development of the game using SlimDX (June 2010) are also very useful, especially regarding Direct2D, which I will most likely use.