Intermezzo Op. 118/2 - Roman-numeral analysis

[allchopin]

As the title suggests, I am looking to do a Roman-numeral analysis on Brahms' well-known intermezzo and I could use some advice.  First of all, would anyone be willing to go through the piece (sans the refrain around m. 76) together and discuss an appropriate analysis?  Second, are there any helpful websites dealing with a strictly Roman-numeral anlysis this specific piece (including figured bass, modulations, suspensions, etc.)?  Thanks for any help. 

First technical question: is the first measure (excluding anacrusis) a constant IV⁶₄ or IV - I with pedal point bass A?

[brahmsian]

You're talking about chords when referring to Roman numerals I assume?

I think the first measure would be IV- I sus4 (technically). The sus4 is wierd because it has the third in there. I don't think it would be a constant IV6/4. You can hear a distinct change in tonality.

[allchopin]

I'm not sure what suspension you're referring to (the D in the bass does not resolve), but I don't quite hear a complete shift in tonality.  He could just as well have written an E like the anacrusis, but did not choose to do this. 

[pianistimo]

there's several articles that i found that might be of some help:

 google: fulltext/Musicologica%206/Musicol6-13.pdf
 click on the article below it (about chromatic mediants)
   (on page 32)

http://depauwform.blogspot.com/2005_03_06_depauwform_archive.html
   (starting:  "this piece is gorgeous...)

other ideas

http://brahms.emu.edu.tr/music/dept/asli-thesis.pdf  (see chpt. 3)

http://orpheus.tamu.edu/couch/Professional/BrahmsOp118No6.pdf
(this last one, i think, has an example at the end of how to analyze brahms largely - instead of chord by chord)