#1
My understanding of how secondary dominants work is that they're all dominant chords, so what works over one should work over another, if you transpose it. For example a line that works over B7 when B7 is the V of ii should also work if B7 is the V of vi. So recently I've been doing just that, taking lines I'd play in one situation and putting it in another situation. But what I've found is that not everything sounds as good in the new context.

For example, when I play <G E7 Am7 B7#5 Em7 Am7 D7> I like to use the whole tone scale over the B chord. Sounds real good. But these lines sound harsh and unpleasant to me when I play something like <D B7#5 Em7 A7> or <Em7 B7#5 Em7 B7#5>

Has anyone here figured this out? Is there some official music school verdict on this stuff?
#2
Brief answer: if resolution chord has 3 in it, use a dom scale, a 4th below, with a nat 6. If resolution chaord has a b3, then replace nat 6 by b6 (or #5).

To an extent, the dominant chord type (and hence usual candidates for melodic material on top) tend to be aware of the upcoming chord.

So, if you have dom resolving up a 4th to a chord with a major triad from its root, then the melodic material over the dominant sounds better with a natural 6th (which is the same pitch as the 3rd in the major triad).

So, A7 => D (or D6, Dmaj7, D7 ...) will usually sound better with an A dominant scale that has a 6 (e.g. A mixolydian, or A half-whole, or A Lydian b7. But also A whole-half).

But if the target chord is a minor (triad, m7 ...), then the melody over the dominant sounds more natural with a b6 (#5), rather than a 6. Because that b6 is the b3 in the minor triad.

e.g A7(#5) ->Dm7. A Mixolydian b6 sounds more natural over the A7#5. For same reason, A whole tone will work (has a #5). Again A whole-half

So, with secondary dominants, you need to consider the target of that dominant.

E.g.

In C maj, if we have Cmaj7 A7 Dm7 E7 Am7 ... the A7 and the E7 (both secondary dominants) could use some kind of dominant scale with a b6, because their targets are both m7 chords.

The next note to cautious of is the choice of 2 or b2 in the dominant scale. To be on the safe side, just see if there is a diatonic chord a b2 or a 2 above it. Use that to guide your choice.

So, in C maj, E7 -> Am7 we known after E is F in C major (a b2 above), hence can use E HM5 (it has both b6 and b2, Mixolydian b6 b2).

But with A7 -> Dm7 ... B follows A in the scale of C, a 2 above. So, could use Mixolydian b6 (i.e a natural 2 in that scale)

cheers, Jerry
Last edited by jerrykramskoy at Dec 4, 2015,
#3
Quote by scarletcantos
My understanding of how secondary dominants work is that they're all dominant chords, so what works over one should work over another, if you transpose it. For example a line that works over B7 when B7 is the V of ii should also work if B7 is the V of vi. So recently I've been doing just that, taking lines I'd play in one situation and putting it in another situation. But what I've found is that not everything sounds as good in the new context.

For example, when I play <G E7 Am7 B7#5 Em7 Am7 D7> I like to use the whole tone scale over the B chord. Sounds real good. But these lines sound harsh and unpleasant to me when I play something like <D B7#5 Em7 A7> or <Em7 B7#5 Em7 B7#5>

Has anyone here figured this out? Is there some official music school verdict on this stuff?
The only thing I can see that might be causing the difference is the preceding chord.
To begin with you're choosing an unusual scale (wholetone). That's more "out" than it needs to be. It may well sound good in certain contexts, but generally this kind of question is all about voice-leading: transitions from chord tone to chord tone, and how you resolve melodic lines.
When B7#5 follows Am7, the G shared tone might help the augmented chord follow more smoothly. (The G is shared with Em7 too of course.)
When it follows D, however, the shared tone is A - so the #5 on the B7 is more of a surprise (not a "bad" change, obviously, just less smooth).

In addition, Am7 > B7#5 includes two chromatic descents (C>A and E>D#) along with the whole step bass rise and the shared tone.
D > B7#5 involves two rising chromatic moves (F#>G, D>D#), with one shared tone and a 3-note bass descent. This is a more tense combination.

So the difference you're hearing may be down to that, and not the phrases you're playing (assuming you're playing the exact same phrases in each context, and voicing the chords the same way).

Likewise, when B7#5 occurs between two Em7 chords, there's (potentially) an odd confusion of voice-leading. The D# on the B acts both as leading tone up to E, and as a descending voice to D. In both cases, we don't normally hear the reverse move directly afterwards. (If B7#5 resolves to Em, there wouldn't usually be a 7th on the Em, unless the next chord contained a C# or C, so the descending voice continued.)
That doesn't make those two alternating chords "bad" necessarily, only unusual. That in itself might make improvising on it tricky to handle.
But again, it comes down to those note-to-note transitions between chords - in your melody and in the harmony. If something doesn't sound right, that's the area to examine: not just the phrases you're playing (and how they sit with the chords), but the way you're voicing the chords.

I'm just guessing of course, because I don't know exactly what you're playing!

I suggest you work with scales closer to the key in question when testing these things out. Use a plain dom7 (or even just a major triad) for each secondary chord, and only alter the scale of the key by one or two notes (the chromatic note(s) in the chord).

If you want to stay with B7, it can be a secondary dominant in five situations (maybe more?):

To Em in key of G. (Scale = E harmonic minor, ie raise the D to D#)
To Em in key of C. (Scale = E harmonic minor, ie raise the D to D# and F to F#)
To Em in key of D. (Scale = E melodic minor, ie raise the D to D#)
To E in key of A. (Scale = E major, ie raise the D to D#)
To E in key of B. (Scale = E major, ie lower the A# to A)

You only actually need a 7th on the B in the last case. In the previous 4, the 7th (A) is diatonic, so makes no difference. (Remember "dominant" just means V, not a 7th chord type.)
And keep the Em and E as triads to begin with.

Of course, you can use other scales if you like, but those represent the least alterations to the scale of the key. When you know how those sound, that's when to start thinking about alterations - one by one - and see how they work in each case. It's all about chromatic resolutions to chord tones.
(The wholetone scale is a cool sound, but it still needs to resolve on to the next chord.)
You can also add diatonic 7ths on the target chords, see what difference they make.
Last edited by jongtr at Dec 4, 2015,
#4
I don't think that's the best approach. Even if "it works" things will sound bland and detached and generic if you just play that way, with that sort of philosophy I find. What you want to do, imo, is focus on phrasing, and coherent melody lines, not "what works". Music isn't "things that work" any more than comedy is "phrases that make sense".

Usually you will want your reference to be the key, not the chord.
May explore progressions later.
Last edited by fingrpikingood at Dec 4, 2015,