Re: jazz chords

[Atte André Jensen]

On Sun, 22 Oct 2000, Jeff Henrikson wrote:

I'd rather not go into detailed discussion about what is better. The
Berklee system is fairly widespread, but that doesn't make it better. It
even seems that Berklee students don't always agree with the Berklee
system.

So let's not argue about systems, there might be a need for more than one
notation system in Lily.

> I have a couple of comments on jazz chords.  This is stuff I have
> thought for a while, but since this is currently being discussed
> I thought I'd mention it.

<snip defs of the Berklee system>

<snip>
> Notate them by any of
> b9, 9, #9, 11, #11 b13, 13 inside parenthesis, 
<snip>

Don't you guys use b10? 

<snip defs of the Berklee system continued>

> That said, I don't think anybody should go changing lilypond to compensate 
> for the Berklee system or any other for that matter, as
> they won't mean anything if nobody uses them.

I think you (being intimate familiar with the berklee system) should
define that standard in lilypond (look at the .scm file James made from
my defs) under the name berklee, and post it here. Then we could all write
"ChordName.Style = berklee" and get chords conforming to your system.

> What I am confused about is why knowing what notes are in the chords is any 
> of lilypond's business. 
<snip "lotta theroy">

Well obviously Lily comes from a classical way of thinking. So she tries
to interpreet whatever you give her and spell it out for you. She thinks
in voicings with actual, exact notes but jazz musicians thinks in more
abstract chords with a function and a quality and a color.

I am gonna make my defs of chords (to be posted here in a few
days...) in this way: What - exactly what and only what - you put in the
chord in Lily (ex: C:7.13-) is reflected in the chord symbol (so
C7(b13)) regardless that most people would put a nine in there also
and probably also leave out the 5th. The "problem" is that Lily
understands voicings and not chordsymbols. I don't know what you would do,
but normally (in a standard-type harmony) I wouldn't even notate any
extensions in my chord symbols. I try to think of the chord symbols as an
outline of the harminic functions, and leave extensions up to the
interpreter, since this depends largely on style, taste and what the
soloist is playing. Of course things are different if you're wearing your
arranging shoes or writing charts for a more complex tune like a Wayne
Shorter tune or in fusion.

I believe it would be possible to make Lily understand something like
"C:7(b5,b13)" since she already understands "dim" and "maj", but basically
you're giving her the same as with "C7:5-.13-", so I don't think that's
worth the extra defs. 

Finally after thinking it through and discussing it with others I think
I'm gonna let my chord-def, inconsist as it might be, reflect real life
notation. For example I discovered that while I write
C"triangle"7(#11) with extensions in () I also write C"triangle"9 with
the extension 9 outside (). And I also write C6/9 and *not* C6(9). So what
I would like to do is reflect the reallife shorthand I write in hand in my
Lily defs.

> Basically, I think we should let lily just do notation. 

Yes!

> Do it right and handle the things that it should, but doesn't (to my
> quite
> limited knowledge- apologies if I am incorrect), such as:
> 
> 
>          G        Fmaj7 /            (b13)
>         ---            / G        F7 (#9 )     D-7  /  / G7 | Cmaj
>          F7
>          ^this by the way is the               This is how to notate
>           more explicit symbol for             rhythm when the melodic
>           the voicing above.  The G            staff is already used
>           means G major triad.                 for a different rhythm.
>                                                In extremely irregular
>                                                cases, "slash"
>                                                noteheads are specified
>                                                just for a measure
> 
> Typset the things as symbols, but don't even pretend to know voicings for 
> chords, unless there is a reason, like somebody thinks a
> really cool hacking project is to get lilypond to look at the melody and 
> chord symbols and generate voicings for two hand piano
> that fit, (and play in rhythm?) or else can analyze a score in notation and 
> lossily reverse construct the chord symbols or
> something.  (Incidentally, there is a whole subject called Berklee chord 
> scale theory that would make a nice framework.)  IMHO
> totally irrelevant for a music typsetting language, but whatever floats your 
> boat.  Just draw the symbols and play silence in MIDI.
> (Also make sure you have a reasonable way for getting the changes to appear 
> only once on the score but printed on every part.)

Well this is with your arranging shoes on. Nice shoes, I just didn't wear
them, which I normally don't...

> Basically, by assuming we know the notes ("voicings", as opposed to "chords") 
> we are just binding our hands to never really handle
> jazz in an even semi-plausable way.

I agree, see above...

Finally I have to mention that you might wanna set your mail-program to
wrap lines after the default 72 chars. Had to bring it up, sorry :-)

-- 
Atte André Jensen