Oh, how I love Brahms. The Intermezzo is in a composite ternary because the A section can be labeled as binary. The A section comes in 4 bar phrases with the first HC in the key of A at measure 4. This phrase is basically repeated, the only difference being an increase of texture, and an IAC at m. 8 in the key of E. The D sharps are the key that we're switching to E. Yep, repeat the period again, HC (in A) at m. 12. As if I couldn't get enough, Johannes gives us one more and a very expected IAC in E at m. 16. A makes up a parallel repeated period (also progressive).
At m. 17 we're back in A and there's an IAC at m. 20. This is our b section. It's tonally ambiguous and lasts for 12 measures. The chords don't seem to make any sense as to declare a key.
At 29, it seems like we're back at the A section (sigh of relief). At first is appears to be the same, but we'll notice that both the melody and the texture (especially in the left hand) is somewhat different. We end on a nice PAC in the key of A Major.
The B section starts at m. 49 and we're in f# minor. After the first 4 measures there is a HC and then a transition to the key of c* major (the dominant) for a PAC in m. 56 and continuing on for a while in the b section. there's a PAC at m. 60 and another at 64. The original melody from the start is b is being used, but has been transposed to the dominant. At 65 we're back in f# and there's a HC at m. 68. I feel like there's either a HC at m. 72 or that this next 8 bars is a transitory phrase to get us back to the recapitulation of the A section.
A starts up again at m. 76. We get the a once, followed by the b, and then with our a again. Obvious binary. I love it. All this is the same as the first time with four bar phrases and what not.
hmm.. i think i'm sleepy now....that means i'm done!!