Re: Additive time signatures

[Hans Aberg]

On 28 Feb 2008, at 23:23, Joseph Wakeling wrote:

> > > I was wondering more if it was possible for that to be a
> > > generally-set preference, along the lines of "\set  
> > > Staff.timeSignature
> > > #'style = note" (that's pseudo code, obviously:-), and also how  
> > > complex
> > > the resulting note could be: for example, imagine a time  
> > > signature along
> > > the lines of (again, in pseudo-notation),
> > >
> > > \time 3/(4.~ 4)
> > >
> > > (Basically this is a rather complicated way of writing a 15/8 time
> > > signature subdivided into 3 beats of 5 eighth notes.)
> >
> > I don't understand this notation; please explain your idea.
>
> The idea for what would _appear_ is a 3 over a dotted quarter note  
> tied
> to a quarter note (hence the notation, which as I said, is pseudo- 
> code).
>  That is, a 3 over a note lasting 5 eighths.
>
> It's not something particularly urgent for me to be able to do, but
> seems like a good example of a weird-but-logical use of the
> number-over-note time signature notation.

OK. I see. Strictly speaking, the time signature is not a ratio, but  
a multiplication. Hindemith does not use a "/" even when writing time  
signatures in line.

In your example
  3 x (5/8)
where (5/8) here is the written duration taken 3 times.

The problem is really to handle subbeaming. Bartok used the time  
signature
  4   2   3
    +   +
  8   8   8
or 9 = 4+2+3. But what is the relative strength of the metric  
accents? - The notation does not tell. In fact, this form of a 9-time  
is played in Romania, and assuming it, it should be:
     >  (>)  -
  9: 4   2   3
so a logical description might be 4+(2+3).

Then one might want to let this be communicated in the beaming. In  
9/16, the first 4 should be one group, and in the 2+3 should have one  
common first level beam, and the second level beam should be broken  
between the 2 and the 3. This will make it different from the  
Bulgarian Daychovo 9 = (2+2)+(2+3).

So then the association between metric description and beaming starts  
to become complicated.


> > We discussed this in the bug list - a typo in Snippets. YOu can see
> > there how it is possible to something in that way.
> >
> > The problem is if one would want to fine tune it. For example, the
> > Bulgarian kopanitsa 11/16 may be described as 4+3+4 where 4 = 2+2. So
> > one may want to typeset the 4 as one group or one group where the
> > secondary beams are grouped as 2+2. Then it is difficult to get the
> > stems in say a 1+1+2 pattern to come out right.
>
> Have you got a link to this discussion?  It sounds interesting.  I'd
> like to understand more of what has been discussed before  
> commenting on
> the problem you raise.

All mailing lists and archives are listed on
  http://lilypond.org/web/about/
The thread I started is on
  http://lists.gnu.org/archive/html/bug-lilypond/2008-02/msg00242.html
which gives a reference to the Snippets manual. Follow the thread; it  
is interesting. In my last post, I give s more complicated example:
  http://lists.gnu.org/archive/html/bug-lilypond/2008-02/msg00242.html

  Hans Åberg