I struggle with the rhythms of rubato-heavy classical pieces, and no one loves rubato more than the Impressionists. When I started listening in earnest to recordings of Debussy’s “Clair de Lune,” I couldn’t even guess the time signature, much less place notes in the bar. This piece is therefore an excellent use case for aural learning through remixing. First I tried putting the MIDI in Ableton over some beats. Then I thought it would sound better to use human performances and cooler beats.
This was my toughest remix challenge yet. Adapting breakbeats to triple meter is one thing; adapting them to 9/8 time is another. I did finally discover that “The Crunge” by Led Zeppelin is also in 9/8, and after some minor editing, the opening drum break fit in just fine. I also used jazz drumming sampled from McCoy Tyner and Adam Makowicz. I used aggressive low-pass filtering to keep the beats from overwhelming the delicate piano, and I beefed up the piano part via compression as well.
You can also hear the rhythms clearly by listening to the MIDI with a simple drum pattern underneath:
“Clair de Lune” seemed unstructured to me on first exposure, but it hangs together more logically than I thought. Here’s a MIDI visualization I made of Alexis Weissenberg‘s recording.
I find it most convenient to think of the piece in terms of pop/rock song structure. Measures 1-8 are the first half of the “verse.”
No wonder the rhythms are hard to hear! Every little melodic figure starts an eighth note behind the beat, and there is a ton of syncopation within the figures too. The piece is in Db major, but right in the second bar there’s a wistful turn to parallel Db minor. At the end of measure six, there’s another wistful turn to relative Bb minor.
Measures 9-14 are the second half of the “verse.”
The syncopation continues as it did in the first eight bars. I hear the Bbm in the last bar as effectively being a Gbmaj7.
Measures 15-26 are the “prechorus,” a long buildup to the hook.
The first four bars are on a floaty Ebm7 chord. Then in the fifth bar, there’s a dramatic chromatic rise in the bassline to the four chords in measures 25 and 26. I hear these chords as acting like an extended Ab7sus to Ab7.
Measures 27-34 are the “chorus”, the hook, the part that gets stuck in your head.
This is the simplest part so far, rhythmically speaking. Everything lands squarely on the beat, though usually people play it with so much rubato that you’d never guess it. The chord progression in the first two bars is a cool one: Db to Fm to E/G#. That’s the tonic, then the iii chord, then the bIII chord in parallel Db/C# minor. The fourth bar gets its angelic lift from the V/V chord, implying Db Lydian mode. The last two bars are a nifty extended Ebm7 chord ending on a suspenseful Ab+.
Measures 35-42 are a variant on the “chorus.”
This section begins identically to the previous “chorus,” but at the end of the second bar, it unexpectedly stays in C-sharp minor for a long time. There are three straight measures of F#m7 over a descending bassline, which is super hip. You can see all the fascinating counterpoint clearly in the MIDI.
Measures 43-50 are another variant on the “chorus” over different harmony.
This is one long Ab7 chord with a few little dips over to Gb. At the end, the Gb6 is followed by Gbm6, the despondent ivm6 chord from parallel Db minor. It’s the Beatles cadence! So sad.
Measures 51-58 are a return to a the first half of the “verse,” though it’s not quite identical to the opening section. It goes to relative Bb minor more quickly, for one thing.
Measures 59-65 are the second half of the second “verse.” This is the first section so far that is an odd number of bars long.
The first four measures use a lot of V/IV. Many writers of program notes have described the effect as “bluesy.” I wouldn’t go that far, but it is a nice effect.
Measures 66-72 are the final “chorus.”
Debussy leaves the melody off the first and third bars, so you can pay attention to the lovely arpeggios underneath. Everyone makes a big deal about how the final Db chord has no D-flat on top of it like you’d expect; Debussy is supposedly leaving it incomplete to express the yearning in his heart or something. You be the judge.
“Clair de Lune” has been orchestrated several different times. The version by Lucien Cailliet is the one that appears in the iconic fountain scene in Ocean’s Eleven.
There have been many other creative arrangements of the piece. My favorite is the one by the San Francisco Saxophone Quartet. (I remixed this one too.)
There are a ton of corny jazz arrangements. I won’t bother to include any of those; instead, I give you Kamasi Washington’s extremely non-corny recording.
Roxane Elfasci plays the piece on guitar. It looks difficult! Though she appears to be playing it in D rather than Db, which helps.
The piece sounds great when sung by a choir.
Since the saxophone quartet works well, it’s unsurprising that a clarinet quartet would work well too.
I also enjoyed this arrangement for bassoon sextet.
While I was writing this, I discovered that Debussy himself cut a piano roll of “Clair de Lune” in 1913.
The fast tempo was probably due to the piano roll’s length restriction. Most performers play slower, which is understandable, but they also lay the rubato on a lot thicker. Debussy obviously wanted rubato (he says so on the score), but he didn’t want as much as people are using. He himself sticks closely to the note values on the page. I mean, you would expect the composer to play what he wrote. I like his take the best; usually performers don’t improve on his hip rhythms.
Bruce Berr distinguishes between “prosaic” rhythm, the “mere timing of events decoded from the printed page,” from “poetic” rhythm, “encompassing virtually everything musical and physical that occurs between, around, and through the beats – curves of energy that ebb and flow in cycles.” Berr confirms a suspicion I’ve always had, that many not-so-good performers of Debussy are ignoring the prosaic rhythms and getting lost in their conception of the poetic rhythms. Berr also doesn’t want students following the prosaic rhythms too closely; it’s a matter of striking a balance. He urges that students “learn how to produce pliable floating performances that don’t necessarily sound ‘rhythmic’ on the surface, yet are pulsing and breathing in accord with Debussy’s meticulous notations.” It seems to me that following the prosaic rhythms more would help with that. But this is a matter of taste. I want to hear the prosaic rhythms. There’s plenty of opportunity for nuance in microrhythm and groove without needing to throw the written note values completely out the window. That’s what black music does. Metronomic time doesn’t have to be rigid or boring; the entire history of jazz, funk and hip-hop proves that.