Re: inversions in lilypond 1.8.2

[Ray Brohinsky]

Han-Wen Nienhuys wrote:


> I didn't give any thought to it. What do you precisely need? How
> should <c e a>  be printed in ChordNames ? 
>
>
> Han-Wen
> (chord agnostic)

Believe me, I was a chord agnostic for years. At least as far as theory 
and voice leading are concerned. Guitars have a nice way of saving one 
from that sort of thing. And then Ken says, Hey, the guitar solo and the 
keyboard pads don't happen at the same time, so why don't you strap on 
the guitar and play the pads as well!!

And now I'm a true-sort-of-believer.

Anyway. The notation already has the ability to move a note in the chord 
to be the bass, ie, a:m is <<a c e>>, and a:m/c moves the c down and 
makes <<c a e>>. This may be useful in jazz, I don't know. But the 
system would easily encompass inversions if using /c identified the 
lowest note in the inversion. So a:m/c would make <<c e a>>  and a:m/e 
would make it <<e a c>>.

As far as what moves down and what moves up, I prefer consistancy: leave 
the a where it is, move the others around it as needed. Then, if I need 
to, I can shift the whole chord by ' or , as needed.

For third inversions, you're forced to have a seventh, which is placed 
at the bottom, and the rest of the triad will stay where it was. For 
root, first and second inversion seventh chords, with the root staying 
put octavaciously-speaking, the seventh will be on top for the root 
(uninverted) form, and nestle under the tonic for the first and second 
(and third) inversions.

What to do about larger chords than the seventh? Well, my theory doesn't 
include larger inversions than three. Someone who does them will have to 
advise, but since an inversion that puts the ninth in the bass is pretty 
weird, and an octave translation of the ninth down just makes a 
good-old-second chord, I doubt there's much argument to be had over it. 
(I'm usually wrong about arguments, though.)

So for me, the ideal would be
root                    a               <<a c e>> or <<a c e bes>>
first inversion         a:/c            <<c, e a>> or <<c, e bes a>>
second inversion        a:/e            <<e, a c>> or <<e, bes a c>>
third inversion         a:/bes          <<bes a c e>>

These are normal for 'closed' form. If someone's getting into open form 
(where there's more than a fifth between any of the upper voices), 
they're going to want to notate that carefully and by hand. In those 
cases, being able to have two streams, one for chord names and one for 
the actual note voicings, makes sense anyway, because it's very hard to 
specify open spaced chords meaningfully in 'inversion' terminology 
without losing the compound intervals.

The syntax for adding a bass note looks just fine to me, btw. Likewise 
the syntax for excluding a note. If making inversion support doesn't 
break that, I'll be very happy. Sometimes I have to play an inverted F 
chord over a G bass (kind of like a heavy sus4) or a D bass against a
C2 chord without the e... I'll bet the jazz guys have a really neat name 
for that! I...just don't have a clue what it'd be.

raybro