Home > Projects > Three Compositions (1995)

My undergraduate thesis at Amherst was titled "Three Compositions". My composition recital was recorded, but the recording is analog and not very good. (Just one year later, in 1996, I got an excellent recording of my Wind Quintet which, being digital, still sounds good today.)

Recently I decided I wanted to get decent recordings of my thesis compositions, so I have been recording them using GarageBand. Hopefully someday I'll get another chance to hear these performed and recorded by live musicians, but these versions will do for now.

Variations on a Theme by Prokofiev

Today he plays jazz; tomorrow he betrays his country.

--- Stalinist slogan in the Soviet Union (1920's)

When I went to Lewis Spratlan junior year and told him I wanted to do a composition thesis, I was just this guy who had taken some theory classes with him. I had never taken any composition classes or really even composed anything. Obviously, he was dubious.

He wrote out the love theme from Prokofiev's Romeo and Juliet ballet and suggested that I write a set of variations on that theme. Then we would discuss whether it was realistic for me to do a composition thesis.

I wrote them during a couple of weeks the summer before senior year. I wanted to write for a small ensemble, but he told me to just write for piano. "But I'm not a pianist." "Well, you're not an orchestra either."

Variation 1 is a retrograde inversion (the melody upside-down and backwards); variation 2 is a straightforward reharmonization; variation 3 is the melody in the form of a blues; variation 4 is an augmentation (one note of the melody is used in each measure of the variation's melody); variation 5 is a more interesting reharmonization.

Sometimes sets of variations are ordered so that the theme can be easily heard in the first ones, with the later ones getting more and more abstract; for this set, that ordering would be Rijival, Gender Traffic, S.C. Junction, F'buta, and Trust Fund. However, I ordered them instead by putting the best two pieces first and last (a trick I learned in the DQ).

Variations 2 and 4 are weak; their main virtue is their brevity. I only included them in the original set because I felt I needed 5 variations. I'm only including them here for the sake of completeness.

However, the other three variations were good enough for Professor Spratlan to OK me writing a thesis. (He also liked that I had transcribed the theme from the record.) For my recital, I arranged them for Big Band, with the help of Andy Jaffe.

Short Pieces for Saxophone Quartet

Look guys. . . I just can't handle these changes. . . I'm not sure if it's the rhythm, or the tempo, or if it's just that I'm a cow.

--- a saxophone-toting cow from Gary Larson's Far Side

MIDI sequencing programs allow you to view and edit music in a "piano roll" view, where the notes are horizontal bars. The higher on the screen, the higher the pitch; the longer the bar, the longer the note. I wrote these pieces mostly by "drawing" notes directly on the screen. This short-circuits all the musical habits you have built up (where you place your hands on the keyboard, how you draw notes on a staff) and gets you closer to writing what simply sounds good to you.

I was nervous showing the first of these to Professor Spratlan, since it was about 45 seconds long. I asked whether I should try to rework it into a longer piece, and he suggested I leave it as it is. "I think it's darling as hell," he said, thus naming the piece.

They are indeed short -- all under 2 minutes -- and difficult to perform. We managed, but the effect was very different from the precise computer-rendered versions here.

Ironically, the one which sounded best live (Pluperfect) sounds awful when rendered with the fake software saxophones. (I rendered it here with a mellower instrument to take the edge off.) The reason is that it's the one song I wrote the old-fashioned way, sitting at a piano. The others were composed using piano-roll, so naturally they are optimized to sound good when listening to playback from the computer.

Aristotle vs. Godzilla

Beauty depends on size as well as symmetry.

--- Aristotle

This was my one "long" piece. Written for 11-piece jazz ensemble (the "Eleventet"), it clocked in at 20 minutes or so. However, the actual music I wrote for it amounted to just a few minutes, the same as the other pieces. It was "padded" to length by improvisation.

The structure was basically three different jazz songs, arranged in the format A-B-C-B-A. The transitions between the sections were handled by four duets. The duets were improvised based on verbal instructions from me (wearing my Charles Mingus hat). For example, the instructions for the last duet looked like this:

1. After the Swirly soli, R.S. drops out. Violin plays Ab blues licks heavy on descending tritones while Trombone plays lots of Eb's and Ab's.
2. Both slow the time down to a medium blues tempo (125-130). Trombone and violin can solo more freely (Ab blues), but both players should keep good time.
3. Shortly thereafter, the violin solo peters out, the trombone starts to sit on a Db, and violin enters as written, in new tempo.

Obviously, it's not possible to render all this with MIDI instruments. The live recording is not bad, but I find it hard to listen to because I personally caused a train wreck in the middle section. The 10 other musicians did amazingly well considering how little rehearsal time we had, but I should have picked someone else as Alto 2.