Re: chord-names

[Jan Nieuwenhuizen]

"Robin Davies" <address@hidden> writes:

> I'm ramping up for a major assault on chord-name.scm in an attempt to get to
> work properly across the board. Fascinating problem. fyi, the general
> approach I'm intending to take is:

Your approach looks well informed, I'm very much looking forward to this.

> 1) separate the existing alists into two separate lists.
[..]
> Hopefully, this simplify the logic considerably and
> reduce the current schizophrenic use/misuse of the alist tables.

Yes.

> As an example, the skeleton of a Maj9 chord is (0 2 0) (0 6 0) (1 2
> 0) to which
> +5/-5 or no5, +11/-11 +13/-13 could be applied as an alteration (in
> American/Jazz, anyway). This is a sensible approach for Jazz and American,
> although less so for Banter, I suspect.

At the risk of repeating myself; I know very little about chord names
apart from the lack of consensus/strong desire for personal/group
taste.

To put an end to endless discussions that arose every time chord names
came up (and that's before there was any implementation), I chose
Banter style, because it was so very logical, and available to me.

I don't think it serves any purpose apart from assuring a customisable
reference implementation.

> 2) Provide a substitution mechanism for "m"/"-" and Maj/M/delta/solid-delta
> within the base chord names.

That would be nice.

> 3) Either noX (banter) or addX (american).

Ok (but not fixed to those styles, I assume).

> 4) A couple of aditional features suggested in the FIXME notes.

Ok.

> 5) Optional support for major/minor/augmented/diminised alterations (my
> personal pet peeve). e.g. +5 (augmented fifth),  but flat-13 (minor 6th or
> 13th),

> 6) Stacked 2xN (Jazz) or "/" sepearated alterations.

What do you mean by this?

> 7) American, Jazz and Banter meta-styles that set the appropriate
> sub-options.

Ok, that's what I would like to see.  I suspect that people will want
to customise their layout anyway (eg, german/jazz, german/american +
exceptions).

> 8) Drop the Jazz and american penchant for ignoring noX (no3, no5, e.g.)
> alterations. If the user doesn't want noX then they should have provided X.

I'm not sure about this.  I seem to remember a request to never show
noX (Amelie, you're still around?).

> 9) The inversion/root ('/' vs '/+') functionality doesn't seem entirely
> sensible to me, but I'll support it as given.  e.g.   a:m7/g   produces
> Am/G in the current system instead of Am7/G, although a:m7/+g does produce
> Am7/G.   a:m7/e produces Am7(no5)/E.

What would you suggest?  I do think we need both, ie, a chord with an
extra added base note (/+), and an inversion (/)?

> 1) German note names in root, and base.
>
> Refresh my memory, please. This means using H for a-flat? I'm not sure
> what's required.

H for B, B for B-flat.  See ly/deutsch-init.ly, ly/german-chords-init.ly.

> 2) Banter conventions.
>
> Not having ready access to the Banter text, I have to infer what I can from
> the code. I'm perfectly comfortable with how to apply alterations to
> American and Jazz chords.
>
> The Banter 11th and 13th chord notations don't seem sensible to me. Is the
> current output correct, or is it broken? Writing C^9/11/13 for C13 seems
> highly undesireable to me. But then, I'm American. ;-)

But you're right nonetheless ;-) Banter is fairly logical; the only
possible exceptions are, well see the banter exception list.

> c:13    :    C^9/11/13      :  C^13
> c:7.9+   :      C^9+      :    C^9+  (not sure).
> c:11.13- :    C^9/11/13-    :   C^13

Yes

> c:13.9-  :   C^9-/11/13     :   C^9-/11/13  ?

C^9-/13

> c:11    : C^9/11      :     C^11
>  c:11+  :  c^9/11+   :     C^11+   ?

Yes.

> c:m7.5-.9   : Cm^5-/maj7/9      :   Cm^5-/9  or Cm^5-/7/9 ?

Cm^5-/9

> c:11.13-    :  C^9/11/13-    :    C^11/13-   or C^13-  ? (like C^9-)

Yes, C^13-

Do these make sense?  I'm always in doubt about the 7 issue.

> I'm guessing: take the highest degree of the chord, and drop all unaltered
> degrees below that.

Yes, that's it.  Also, add noX for missing degrees (where an altered
degree counts as present too).  But anything hairy and Banter specific
has no priority.

> The current implementation seems to apply that rule up
> to the 9th degree, but provides all degrees for 11 and 13 chords.

And is thus broken.

Greetings,
Jan.

-- 
Jan Nieuwenhuizen <address@hidden> | GNU LilyPond - The music typesetter
http://www.xs4all.nl/~jantien       | http://www.lilypond.org