The Girl From Ipanema - Analyzing the bridge

[princeplanet]

Girl From Ipanema. The bridge - how do you call it? What key are you in for each chord?

[Drumbler]

Girl From Ipanema. The bridge - how do you call it? What key are you in for each chord?

One way is to use arpeggios as the basis of your improvisation. Not think keys.

[JohnW400]

There's a big thread about this here:

'Girl from Ipanema' / basics of analysis - Jazz Bulletin Board

[M-ster]

Agree with Drumbler. I'll add that the chord-scale approach would look like this:

Gbmaj7 |      | B7   |      |
Gb maj          F# mel minor

F#-7   |      | D7   |      |
F# dor          A mel minor

G-7    |      | Eb7  |      |
G dor           Bb mel minor

A-7    | D7   | G-7  | C7   |
A phryg         G dor
         Eb mel minor  Db mel minor

One could argue that the initial Gbmaj7 could be Gb lydian, since the tune has been in F major up to this point, but to me, this harmony change sounds like it's going to amount to a modulation up a half step, when you first hear it -- i.e., oh, now we're in Gb major.

[M-ster]

You could also greatly simplify the first 12 bars by thinking like this (it makes each of the 4-bar phrases parallel):

Gbmaj7 |      | B7   |      |
F# lyd          F# mel minor
 
F#-7   |      | D7   |      |
A lyd           A mel minor
 
G-7    |      | Eb7  |      |
Bb lyd          Bb mel minor

And, sorry JohnW, like the OP, I didn't search first, either. There's likely a pre-existing thread for any of these types of questions!

[ksjazzguitar]

I think of that GbMaj7 as a N6 in F - it just never resolves, which helps explain the #4. In classical sometimes they call that "backwards resolving" - where the resolution comes first.

Peace,
Kevin

[randalljazz]

There's a big thread about this here:

'Girl from Ipanema' / basics of analysis - Jazz Bulletin Board

"+1"

seek out dr. byrne's excellent analysis...

[ksjazzguitar]

I guess that I treat the D7 and Eb7 as backwards resolving too since I treat those as lyd dom. If they were functioning as dominants, then I could use the altered scales, but that doesn't sound right to my ear.

So, to me, that D7 is actuall paired with the F#m7, not the Gm7 - I just don't hear it resolving that way. To my ear it is a bVI7. Note that it's 3rd is the root of the F#m7, just like an Aug6 chord would be. The same thing happens with the next two chords, but the Eb7 serves double duty as it also resolves to the D7 (through the interpolated Am7.)

We don't often use backwards resolution in jazz, but it is common enough in classical, and Jobim uses classical ideas from time to time. I just don't here that D7 or Eb7 functioning as a true dominant in a tonic dominant-relationship and I think that the scale choices support that.

Peace,
Kevin

[princeplanet]

Yeah, checked out the other site's analyses (very interesting) and posted this:


Jerry Engelbach:


The Girl from Ipanema

A ||: FΔ - - - | - - - - | G7 - - - | - - - - |

| Gm7 - - - | Gb7 - - - | FΔ - - - | (Gb7) - - - :||

B || GbΔ (or Ebm7) - - - | - - - - | B7 - - - | - - - - |

| F#m7 - - - | - - - - | D7 - - - | - - - - |

| Gm7 - - - | - - - - | Eb7 - - - | - - - - |

| Am7 - - - | D7#11 - - - | Gm7 - - - | C7#11 - - - || to A section

There are some analyses further back on the thread that are closer to the above changes than what you interpret as V-I's in the bridge. The three keys suggested in the bridge — Db, E, F — are in bold.

B || [Db] GbΔ (or Ebm7) - - - | - - - - | [E] B7 - - - | - - - - |

| F#m7 - - - | - - - - | [F] D7 - - - | - - - - |

| Gm7 - - - | - - - - | Eb7 - - - | - - - - |

| Am7 - - - | D7#11 - - - | Gm7 - - - | C7#11 - - - ||


my response:

I suppose every chord in the bridge up til the Am7 is it's own little island and it's all very arbitrary to try to force an analysis. For a CST approach I hear the Db, E, and F keys, but I don't hear the B7 as the start of the E section, curiously enough. Over the Gb maj7, I try to infer a Db 6 /9 to B7, and following, E 6/9 to D7, then F 6/9 to Eb7 .... ie, a I maj to bVII7 sequence up a min 3rd then up a semitone. Further I hear it as modal interchange, for me specifically when improvising I like to sometimes use Db maj blues for the first chord then Db minor blues for the second chord followed by E maj blues going to E minor blues for the next 2 chords etc. Pretty simplistic I know, but works for me anyhow...

[ksjazzguitar]

...I suppose every chord in the bridge up til the Am7 is it's own little island and it's all very arbitrary to try to force an analysis...

Again, I think that they are related if you look beyond standard jazz cliches. Jobim was a master of taking typical harmonic patterns and turning them on their ears so they are unrecognizable at first glance.

To me the point of the bridge is that the first twelve bars "wind up" and the next four "wind down." The winding down is simple enough - just a simple iii-V-ii-V. The winding up is accomplished by essentially hinting at going backwards through standard chord relationships.

That GbMaj7 back-relates to the FMAj7 as an Neapolitan 6th (Sorry, I erroneously used the incorrect term "backwards resolving" but have since looked it up to double check.) But that GbMaj7 is also lydian so it can be thought of as a IV chord. If so, then that B7 can be thought of as a backdoor dominant to the imaginary I, so that F#m7 is also the ii to that backdoor turnaround (again, going though them backwards to obscure function.) The F#m7 also has its own character as a i to which the D7 functions as a bVI7 a pretty common borrowed chord in minor (that can also be thought of as an Augmented 6th chord.) That pattern repeats up a m2 - again a backwards relationship, we tend to think of things descending m2 more often than ascending. Again, that D7 seems to resolve to the Gm, but the fact that we can't really alter it speaks against it functioning as a V. And lastly, that Eb7 (the peak of our ascent) resolves to the D7 with the interpolation of the Am7, which begins the descent.

So, I see the first 12 bars as an ascent to through fragmented chordal sequences with hazy (but not arbitrary) harmonic relationships. I don't think that you could accomplish this with random chords. This fragmentation of harmonic relationships is a very common 20th century technique - the Impressionists did it with harmony and Stravinsky did it with melody - among others.

I think that things will seem arbitrary if you try to look at it through ii-V-I lenses, but if you are willing to look at it with a more sophisticated harmonic vocabulary, you'll see that this is a very well crafted sequence of chords trying to accomplish a very specific goal.

The ratios are interesting too. I've often heard the platitude that "All great art achieves an intensity peak around 75-90% through and then resolves (a hasty generalization, but one that seems to often hold up.) Isn't it interesting that for the first 75% we have this ascending, quasi-functional harmony that peaks in intensity and then begins a descent and resolution for the last 25%.

Again, Jobim is the master of making simple chords seem complex. I think that if we assume that anything that he does is arbitrary, then we do him a disservice. True, he is using an harmonic vocabulary that goes beyond what most jazzers do, but that does not make it arbitrary - we just need to learn some more sophisticated harmony. He may have stepped beyond jazzes preoccupation with ii-V-I, but there is a lot more to music than that. I know that jazz players like to think that jazz has the hippest chord changes, but there are still plenty of tricks that classical can teach us. Jobim was hip enough to keep his ears open to other possibilities.

Peace,
Kevin

[JohnW400]

You could also greatly simplify the first 12 bars by thinking like this (it makes each of the 4-bar phrases parallel):

Gbmaj7 |      | B7   |      |
F# lyd          F# mel minor
 
F#-7   |      | D7   |      |
A lyd           A mel minor
 
G-7    |      | Eb7  |      |
Bb lyd          Bb mel minor

And, sorry JohnW, like the OP, I didn't search first, either. There's likely a pre-existing thread for any of these types of questions!

Don't misunderstand. There's a lot of information at that AAJ post that might not make it into here. For example Vic Juris mentions something about all the NYC Brazilian players using Ebmi7 where the RB has Gbma7. Vic isn't a member here so that would never make it to this site. I post it only as additional spots to look at. not to derail or take th eplace of posts here.

[docbop]

"+1"

seek out dr. byrne's excellent analysis...

Have you checked out Ed Byrne's books, if so what do you think of them?

[ksjazzguitar]

For example Vic Juris mentions something about all the NYC Brazilian players using Ebmi7 where the RB has Gbma7.

Interesting. I think that actually supports my back-relating bVI7 theory - that would make all of the first 6 chords analyzable as a sequence of i->bVI7, in three different ascending keys. Again, far from "arbitrary."

I think that the notion that that might really be an Ebm7 is supported by that fact that that would make the melody the exact same extensions over each minor chord in mm. 1, 3, and 5 - pretty compelling argument. I fits in with my idea that this is just an intensity building ascent before a resolving descent. Maybe jazzers subbed the GbM7 since that N6 sound was so popular in the 60s.

Thanks, I'd never looked at it that way.

Peace,
Kevin

[M-ster]

... Vic Juris mentions something about all the NYC Brazilian players using Ebmi7 where the RB has Gbma7 ...

That would also make each of the first three 4-bar phrases parallel (in root motion as well as chord-scale).

[ksjazzguitar]

Yes, it just makes it a 4 bar melodic and harmonic sequence (i-bVI7) that repeated in three different keys. Nice. Very tidy it turns out.

Peace,
Kevin

[Reg]

Here's a quick description of Bridge of Girl From Ipanema... check it out and ask anything you want... Best Reg

[randalljazz]

Have you checked out Ed Byrne's books, if so what do you think of them?

have not purchased his books, but am familiar with his approach from a long relationship on "another board".

whether his methods, which involve a lot of work, are right for you, i could not say.

however, his jazz credentials are solid, having played and/or recorded with the likes of chet baker, joe henderson, buddy rich, etc, and earned a doctorate at new england conservatory.

check out his website...sample lessons are available free.

Byrne Jazz - Home

[Aristotle]

I think there too much attempt to connect and rationalize things here. Does it matter whether GbD has a supposed “relationship” to the FD. Same chord type, half-step away, isn’t that generally as unrelated as possible?

There are two obvious paths.
Byrne on the other forum points out that the B-section melody-line is a repetition of:
mi, fa, mi, re, mi, re, do, re
and that would lead one to conclude that the Gb is Lydian, the F#m9 and the Gm9 as Dorian, and 7ths that follow them as Lydian.

That’s not wrong, but in my first view of the song and the one that Reg uses, the melody-line is a repetition of:
ti, do, ti, la, ti la, sol, la
This would lead one to view the Gb as Ionian, the F#m9 and Gm9 as Aeolian, and the 7ths that follow them as Mixo.
And the I-IV7 logic that Reg applied to Gb-B7 would apply to the other pairs, because F#m Aeolian is A Ionian to D Mixo.

I think both of these work easily because they don’t conflict the original melody; and that the Byrne approach is slightly more “outside.”
If my head is working right, in either interpretation, the second chord-scale keeps 5 notes of the original scale and changes 2.

[ksjazzguitar]

I think there too much attempt to connect and rationalize things here. Does it matter whether GbD has a supposed “relationship” to the FD. Same chord type, half-step away, isn’t that generally as unrelated as possible?

Well, that is a common chord sequence, like in i->bVI7->V7. It also can be seen as a chord change as common tone. You could also see it as a sub for m7 chords moving up a m3 to a m6, a very common movement for chords. (Gm7->Bbm6 becomes Gm7->Eb9)

I agree that it isn't functional in this case (since it lacks resolution) but it is related. And the non-functionality is ultimately irrelevant since it is a chromatic sequence. Chromatic sequences do not require functional harmony. It is the parallel chromaticism that gives it structure, not functional harmony. In my experience, parallel chromatic sequences often don't have functional harmony.

To reiterate the case that the first 12 bars are just a chromatic sequence... Based on JohnW400's comment that Brazilian players play that first chord as Ebm7, then we have three 4 bar sections that are exact chromatic transpositions of each other.


------------------
-6-7-6-4-6-4-2-4--
-6-----------2----
-4-----------1----
-6-----------2----
------------------


So, that is the first iteration of the sequence. Now for the next 4 bars, it is an exact chromatic transposition up a m3, not just in chord voicing but in melody. Then for the next 4 bars it is an exact chromatic transposition up a m2. So, we end up with:


-----------------|------------------|---------------------
-6-7-6-4-6-4-2-4-|-9-10-9-7-9-7-5-7-|-10-11-10-8-10-8-6-8-
-6-----------2---|-9------------5---|-10--------------6---
-4-----------1---|-7------------4---|-9---------------5---
-6-----------2---|-9------------5---|-10--------------6---
-----------------|------------------|---------------------



That is not a coincidence. It is clear (to me at least) that that is the unifying element of the first 12 bars here. We don't do this type of chromatic sequence in jazz as much (not to this level anyway) but it is very common in 20th century classical composition and Jobim no doubt listened to a lot of that.

Again, I think that it cannot be a coincidence that there is an exact chromatic sequence repeated 3 times. (My teacher always said that a sequence should be be repeated 2 and a half to 3 times - no more no less.) What would be the odds of arriving at these perfect chromatic transpositions at random? And these chromatic sequences never need functional harmony to connect the iterations of the sequence and don't need much within the sequence - they have a logic all their own, the chromatic parallelism giving it it's own logic. True, someone can point out that the intervals of transposition seem arbitrary (m3 and m2) but they often don't have a logic in these chromatic transpositions. I wish I could get my web cam working, this is so much easier to show. But if you play through the tab, you'll see what I mean.

As to the GbMaj7, of course we can see that as a diatonic sub for the Ebm7. But because of the Dorian nature of this chord, we expect C in there, not a Cb. And I would expect that anyway since it has a back-relationship to the FMaj7 - that C is still in our ear. I don't think that you can build a logical argument for a Cb (yours seemed like a non-sequitor to me, but maybe I was misunderstanding), but if that is what your ear hears that's cool. It wouldn't be the first time Reg hears an 11 over a chord where I hear a #11 - that's cool.

As to the dual function of the D7 (what I might call a pivot chord) - I used to think this way, but have been drifting away from that and the revelation of the chromatic sequence puts the nail in the coffin as far as I'm concerned. Further, it just doesn't sound like it has dominant function to my ears. If the D7 were truly a dominant function to the Gm7 then we should be able to alter it, but to my ears that sounds bad and further it destroys the chromatic sequence - in the chromatic sequence each chord in the sequence should have the corresponding chord/scale. And again, the nature of the chromatic sequence is that you don't need to worry about the relationship of the last chord of one sequence and the first chord of the next sequence - it is irrelevant (or at least not important.) To my ear, that D7 (as the B7 and Eb7) sounds intractably lyd dom, again what we would expect for a bVI7. Perhaps you could build a dual function argument for the Eb7 as the bVI7 of Gm and the bII7 of upcoming D7 - at least those both have the same chord/scale and can resolve similarly.

I still like my idea of winding up for 12 bars before we releasing for the next 4. In light of the new, clear argument for the chromatic sequences, I'm backing off a little on the "back-relation" argument since it is no longer really needed. Of course, the final 4 bars are just a diatonic sequence. So we build up tension for 12 bars, ascending with chromatic sequences, and then descends in 4 bars (having doubled the harmonic speed) with a diatonic sequence that resolves us gently back to the tonic.

That's how I hear it at least. Masterful chord changes.

I have more, but people complain when I go on too long.

Peace,
Kevin

[M-ster]

I think there too much attempt to connect and rationalize things here ...

And then you continue with your own theoretical discussion that includes solfège, of all things.

Thanks for the video, Reg'. Those are interesting contributions to the threads.

[Aristotle]

solfège, of all things.

That's kiddie stuff, or it was back when 'jazz minor' was lydian b7. Ah, progress!

[Aristotle]

Chromatic sequences do not require functional harmony.

Don't tell me, I am not the one trying to explain the Gb as an N6.

Based on JohnW400's comment that Brazilian players play that first chord as Ebm7

I play Ebm9, too.

[ksjazzguitar]

"Originally Posted by ksjazzguitar
Chromatic sequences do not require functional harmony."
Don't tell me, I am not the one trying to explain the Gb as an N6.

I said that they don't require functional harmony. I didn't say that it was forbidden. And officially we aren't in the sequence until the second iteration begins.

I've just heard it as an N6 and that is what I was taught by at least 3 professors - and that's what it undeniably is in relation to the FMaj7 - that is its mathematical relationship. You can say that it has no relation to the FMaj7, but if it has one, it is as an N6. You can call it a pivot chord or dual function, but then it still is an N6 relative to the FMaj7. Again, perhaps because I hear it with a #11 it is easier for me to hear it as an N6 - if you hear it with an 11, then it's going to sound like a tonic so it would just be a direct modulation from F.

But of course that became kind of moot with the revelation of the Ebm7 and three perfectly formed, perfectly executed chromatic sequences. I really don't see how people can ignore that as the most logical interpretation of what Jobim was thinking - as if this was just some random thing, a one in a billion shot. No, Jobim was a sophisticated harmonist with extensive classical and jazz knowledge and he knew exactly what he was doing.

Of course, what Jobim was thinking and how to solo over it are two different things. But I think that understanding how he was thinking will inform any of the options. What the composer was thinking is always a good place to start (but probably should restrain you.)

Sorry, but you are not going to break my high. I have tried for years to figure out what he was thinking and with that seed crystal of the "Ebm7" I finally do! This is amazing. Like so much of Jobim, it seems so insrutible, but then once you see it it seems so obvious. If some of you choose to not acknowledge the obvious connection with 20th century classical harmonic practice, that is your choice. But you aren't going to bring me down baby!

Peace,
Kevin

[Reg]

Here's some more BS... the tune is not complicated... when I have to play I go through all the possibilities and more which I imagine you would do to keep the tune somewhat interesting. As I said on the video if you want something a little more advanced, I'll post "Ambleside"... I saw chart at gig and it took me a few times through to get it harmonically, the melody is simple, but cool. I had about 30 sec. was fun... best Reg

[ksjazzguitar]

Here's some more BS... the tune is not complicated...

Yes, I know that your interest in theory ends once the question "What scales can I use?" is answered. That's cool. But some of us like looking deeper. The question was about analysis and for those of us who like looking deeper there is some really hip sh!* and deeper harmonic structures going on here that extends well beyond the choice of scales. Some of us find it fascinating. You had your fun now let us have ours.

Peace,
Kevin