Re: square note heads for cluster

[Mark Polesky]

Anders,

> Here's a start on a cluster-chord function.

Exciting!

I don't have the answers to your function-coding questions,
but I thought I'd share a little more of my work with 
generating basic clusters manually. My original idea for a
cluster function was to automate this mess, but I had to
start somewhere.

Some issues (number 7 is the real pain):

1. There are more cluster-styles out there. I wish I had
   access to the Kurt Stone and Gardner Read books, and at
   the moment, there are a lot of composers who've used
   keyboard clusters, and I don't own their scores: Crumb,
   Stockhausen, Berio(?), Cage(?), and Ligeti (in the organ
   works).

2. if the function finds notes INSIDE the chord that are not
   either of the 2 outer notes, it should at the very least
   suppress the printing of the accidentals. Even an 
   unaltered note within the chord may be given a courtesy
   accidental, which would likely be unwanted. 

3. I had intended to have my function remove any given inner
   notes, and then add ALL the natural (ie. not sharp or flat)
   notes between the the 2 outer notes, so that a new stencil 
   is drawn for each note. I haven't studied your function to
   see if that is what you are doing.

4. clusters with different outer- and inner-note-heads
   (like "full-width" "Cowell" and "Sciarrino") are tricky
   for half- and whole-notes because the inner-note-head
   stencil must not cover the inside of the half/whole-note.
   Experience has shown that 2 inner-note-head stencils are
   needed in these cases: 1 that I called "OuterLink" that
   can sit directly above or below an outer-note-head 
   (these tend to be shaped like bow-ties); and another I
   called "InnerLink" which is usually more like a rectangle.

5. some cluster-styles (like "Sciarrino", for example) align
   the cluster on the stem-side, so the function would need
   to detect the stem-direction, and adjust the stencil
   accordingly.

6. some cluster-styles (like "Cowell", for example) have 
   stems which only connect to one note. I've solved the
   problem manually here, but the problem will ultimately
   have to be solved dynamically.

7. the appropriate cluster-width for "full-width" depends
   on a lot of different factors. First of all, quarter-
   notes, half-notes and whole-notes have their own unique
   note-head-width (1.3125, 1.39, 1.98 respectively, at the
   default staff-size, according to my own measurements), 
   and secondly, there are 8 different font-sizes for the 
   Feta font, and the note-head-widths are not scaled 
   evenly (this is intentional). I don't know how to 
   dynamically find the current staff-size. And for what 
   it's worth, the "font-size" and "staff-space" properties 
   may also affect width (and height, as well), and I don't 
   know how to detect these dynamically either. Finally, 
   all of this may affect other styles (like "Cowell" and 
"Sciarrino")...

Well, the way I've conceived it, this is a big project. I've
been meaning to tackle it myself, but a little bit at a time.
Seeing your enthusiasm has prodded me to share my unfinished
work, however, and if you're up to it, feel free to take
advantage of the work I've done so far.

I've attached my work to date on the quarter-note styles,
designed for use with the default staff-size (20), font-size
and staff-space.

I've done some work on half- and whole-notes, but I'll leave
that for another post.

Hope this helps!
- Mark

clusters-q.ly
Description: Text Data

clusters-q.png
Description: PNG image