Satin Doll Analysis

*Disclaimer: Schenker was a terrible human. For people who might not know, I was influenced by Schenkerian Analysis in that I like see how chord structures might be more or less surface level or used prolongationally, but I don’t have any kind of insistence on a certain background structure, I don’t really look at melodic diminutions, and I certainly don’t use analysis to claim any piece is related to “genius” dead composers or “therefore this music is good/bad.” I tend to call my type of analysis Reductive, because it looks at reducing surface level chords to more structural motions.

I certainly don’t think that all music must have these properties, or that music that doesn’t isn’t good or interesting, it just happens that I like chords! and this is the type of thing that interests me based on my background. I’m trying to branch out, so I chose a bunch of pieces based on things like texture and structure and the composer’s background. I then stuck them in chronological order. It just happens that the second piece in my analysis series is very chordal! So this analysis will have more chords than maybe some of the other pieces on my list. (Even when I do use chords, I’m not going to assume that the chords that are typical in Western Art Music are going to have the same function.)

This week’s piece is Satin Doll is by Duke Ellington from 1953. Here’s a recording:

A ton of what I heard in this piece reinforces different tension builds/relaxes in a similar pattern, where the largest feelings/tension are in the middle. The large form, parts of the small form, other details, give a fairly quick tension build with a slowish let-down.

The overall form of this piece is Intro, Tune, Solo, Tune (short), Coda. Within that larger structure the Tune has an AABA thematic/melodic/chordal structure (each letter representing an 8 bar phrase), and the Solo is loosely BA, and then the second Tune iteration is just the last A.

A visual representation might look like this:
Intro Tune Solo Tune Coda
(a) A A B A B A A (a)
xxy zz’

There are texture changes to emphasize the large form; The Intro starts with just the piano and bass with a definite feeling of anticipation, the whole band enters for A, the instruments that are playing the melody change at B and increase the texture and volume on the internal repeat. This leads to the return of A, which has extra embellishments by non-melody instruments (BLAT!). There’s huge drop in texture for the solo, which is a piano-bass dialogue, the bass does a short solo transition back to A, and then the Coda mirrors the Intro.

Nesting even further, the internal structure of A is what the Music Theory world likes to call Sentence Structure, where the phrase is xx’y (y being twice as long as x). Here, x is a two part melodic motive with associated chords, and the second x is shifted up a step. The y idea is a continuation of the last half of x, moving it further up and then relaxing down into the cadence. For this piece, we can say that upward pitch motion tends to increase tension, and downward pitch motion tends to relax it. Both the Intro and Coda are based on rhythmic ideas from the end of x as well.

The internal structure of B is a phrase (z) that is repeated up a step in pitch. It resembles a classical parallel period, because the first phrase feels more open than the second, but the melodic/rhythmic ideas are the same for both phrases. Some may not feel 2 phrases within B, since the standard phrase length for this piece is 8 bars not 4. To me, the embellishing hits and the long note make it feel like a cadence, even if it is open. Also, because the texture changes for the repetition of z, it feels like a more structural phrase, rather than just a melodic repetition within a phrase.

The high point in tension at the end of B leads to a feeling of coming home when the next A phrase returns. It has extra embellishments, including some really low pedal tones (notes that don’t change much), which increase the range used, and contrasts the very high pitch at the end of the B phrase. This also follows the idea that high pitch is more tension, and lower pitches are less. The tension in this A is less, and the very light texture for the solo (piano, bass, drums only) brings down the energy again. Ellington builds tension with pitch, rhythm, and volume during the solo thru a B section and an A, building to almost the end of his solo, but then the tension falls before building a little during the bass retransition.

The return of A after the solo is not as high tension as the end of B, but it brings back mild anticipation for how the end is going to go. The feeling of anticipation in the Coda is transmuted (from the feeling in the Intro) into a more stable sense of closure, partially due to repetition, but also because the ending changes to end on Tonic.

Both the phrase structure and the overall form have tension and pitch high points near the center, reinforced by texture and melody. That’s a fairly basic overview, and gives a more or less complete picture of the piece, I hope. You can stop there if you want – you can mostly describe the tension and excitement of this piece in terms of texture and form, with some light mention of motive and pitch level thereof. But because it’s me, we’re going to do chords!!

This type of jazz actually works pretty well with Functional Analysis, because it is largely based on falling fifth movements, so labeling things with Dominant and Tonic still makes a ton of sense. This piece also has some nice nested D-T motions on multiple levels. I originally used this piece in my dissertation as an example of a non-V dominant, the tritone substitution chord, which does occur, but I don’t think it’s hugely structural.

Here’s the chords, and some measure numbers for me to refer to:
phrase A m2 3 4 5 6 7
||: Dm7 G7 |Dm7 G7| Em7 A7 |Em7 A7 |Am7 D7 |Abm7 Db7 |CM7 (Dm7| Em7 A7) :||
phrase B m2 3 4 5 6 7 8
Gm7 C7| Gm7 C7| FM7| FM7| Am7 D7| Am7 D7 |G7 Dm7 |G7

[The parentheses indicate that depending on which A section it is, the band might play those chords to get to the repeat of the A section, or not play them if they’re going on to a B section. ]

The bolded chord before the parentheses is the chord where the piece eventually ends. So, we’ll call C Tonic. Typically, the Dominant would then be G, and the most likely PreDominants would be D minor or F, with strong preference to Dm, since this is jazz, not blues. We do see a lot of G and D chords (and they sound right in context), so I’ll use abbreviations: D (dominant) for G chords, and PD for D minor chords, and similar abbreviations for chords at transposed pitch levels.

If you know a little bit about note reading or chords, you might notice that many of those chords are a fifth apart. Measures A1-6 all have internal falling fifth motions, but not between the measures. Measures B1-2, 5-6, 7 all have internal falling fifths too, and there’s falling fifths over the barline from B2-3 and 6-7.

With Functional Analysis, the A section looks like this:
1-2 3-4 5 6 7 8
PD D (x2) PD D (x2) PD D PD D T (t PD D)
of C of D of G of Gb C of D
On a zoomed out level, the chord structure also follows a falling fifth D-G- C motion, with the Gb interpolated to provide a close-up dominant feel, using half steps from Db to C and Ab to G to move the progression to the cadence, substituting for the standard G7-C. The turn around parenthesis chords use a falling fifth motion, with both the PreDominant and the Dominant of D, which is the first chord at the beginning of A, so the repeat turn around is very smooth.

The Functional Analysis of the B section looks like this:
1 2 3-4 5 6 7 8
of F of G of C
The last two chords are really there to reinforce the cadence, and help transition from the ending on G back to the C major A section – changing the feel of G as local tonic to definitely dominant of C. But since cadential rest point of the B section is m7 G, the whole Tune structure of AABA ends up looking like CCGC on a really zoomed out level. Even the Intro and Coda are G based, so overall the primary pitch zones of the piece at the most zoomed out level are
Intro Tune Solo tune Coda
and the highest tension point at the end of B is on the G, dominant. Now, I don’t know if I’d call any of these modulations or even full tonicizations, but that’s my read on the chord that is the most important/has influence over each part.

This is why the Db substitution for Dominant is really interesting, because on a largescale level this piece is really fifth based, but the chords at most of the local cadences (since there are way more A sections than B sections) aren’t actually a falling fifth. Structurally, the dominant-tonic motion is a falling fifth, but more locally, half-step motions signal the cadence. Interestingly, in the recording the bass player doesn’t go Db-C in sequence, but adds another leaped note in between. Additionally, the G chords, especially the structural/cadential ones, really feel like dominants, overall I feel an expectation of eventual resolution down the fifth ladder to land on C eventually, but G chords never do in the moment, except at the very end. Throughout the piece, there is falling fifth movement driving local chord progressions, but the expectation of cadential fifth movement is frustrated til the final two chords.

This chordal tension sometimes works with the texture/melody tension from the first half of the analysis, but never getting the dominant-tonic motion completed til the very end provides a continuing chord progression motivation that is somewhat independent from the tension provided by the texture and melody.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s