Analysis?

I love to write music (as do most of you guys, I'm sure...thats why we are here right?). But, the one thing I know I do not do enough of, is analysis. I think that this is a must do for any composer, but I just do not seem to do this enough.

I was wondering what everyone thinks are some great works that SHOULD be analyzed.

First of all, analysis is not very clearly defined. Mostly, it depends on what particularly you want to analyse in a piece, and depending on this question some pieces might have more to offer than others, and some analytical questions might be much more difficult than others.

You can probably grasp the major formal principles of most music within a few minutes, which already constitutes a mini-analysis. Or you can go in the detail and analyse a piece for half a year, then write a 50 page essay on it. I agree with you that analysis is something very helpful for a composer and should not be neglected. The question here is mainly how many of which kinds of analyses you want to do. It's probably most helpful if you mix it up a bit: Do some short and quick general formal analyses of piano pieces or songs (because that helps you to concentrate on form/harmony/etc. without having to read through a huge score). Some "analysis classics" there would be Beethoven piano sonatas (particularly first movements) (mostly for form), Bach inventions and fugues, Schumann piano pieces and songs (or Schubert, Brahms, etc.) (mostly for harmony), etc. You notice that this list is very German-heavy, which is probably due to the fact that I live in a German speaking country and the music institutions here are still quite rooted in this sort of musical tradition :P

This sort of pieces are well-suited to give you your basics in traditional European harmony and form analysis, and are reasonably short and doable, while still holding quite many challenges, even for trained music theorists.

You can of course also do these pieces more in depth and really interprete every single note to the last detail, but personally I'd go to less over-analysed pieces for this. If you really are going to spend a couple of weeks or months analysing a piece, it might be a lot more fun to do it on a piece where you actually might find out things that others haven't seen in it before, which is the easier, the fewer people have analysed it. (If you spend half a year on a piece to find out what others have already written books about, it might be a little discouraging, but it's of course a personal thing.)

I don't know of any piece that you absolutely must have analysed. Do some of those little standard analyses, but don't spend too much time on them, and do some (even just a single one will teach you tons) larger analyses of more "special" pieces of your own choice. Both the ability to quickly grasp the general structures of a piece and the ability to really go to the detail in any piece, no matter how weird and untraditional it is, should be trained by a composer in my opinion.

One particular recommendation I have for every composer: Analyse an electroacoustical piece. It doesn't matter whether you're actually interested in electroacoustical composition yourself. But analysing an electroacoustical piece teaches you some things that you might not learn as directly in any other analysis: First of all, it forces you to listen. There are usually no scores for these pieces, so all you can go by is how it sounds. (Analyses of traditionally notated music sometimes have the problem that the writer concentrated almost only on the written notes, without even thinking about the sound. You don't get this option in electroacoustical music.) Another thing analysing electroacoustical music teaches you is not just to stick to standard procedures when analysing, but developing your own tools and your own language in respect to what you're analysing. Doing a harmonic analysis on a piece of musique concr