Re: triplet brackets

[Erik Sandberg]

On Tuesday 17 September 2002 01.51, Han-Wen Nienhuys wrote:
> address@hidden writes:
> > On Friday 13 September 2002 11.34, Mats Bengtsson wrote:
> > > The default in Lilypond to add a triplet bracket as
> > > soon as you don't have a beam with exactly the same
> > > length as the triplet.
> >
> > I know, the question was whether this should be considered a good
> > behaviour, or if it would be better to have a different default (for the
> > todo). E.g., I can find no example of sane notes where it would be bad to
> > have a rule like "make the brackets invisible iff the bracketed notes are
> > all in the same beam", but I might be wrong.
>
> Detecting all cases where the bracket is not needed is difficult. For
> example
>
>   [c16 c \times 2/3 { c c c } c16]
>
> I'd say it is not very clear which notes belong to the tuplet.

OK, thanks, I can see the problem. However, I still think it's still possible 
to detect the normal case when the bracket isn't needed:

AFAIK, the only normal bracket removal that is not covered now is when a 
group of triplets, all with BeamCount=n, is beamed together with notes with 
BeamCount<n. So it could be solved just by removing tuplet brackets if, say, 
the first tuplet's stemLeftBeamCount and the last tuplet's stemRightBeamCount 
are both strictly less than each other stem*BeamCount of all the tuplets. Or 
equivalently, there exists a m>=0 so that if m would be subtracted from all 
stem*BeamCount of the beamed notes, the rule you use now would remove the 
tuplet bracket.

E.g.:

[c8 \times 2/3 {c16 c16. c32} c8] No bracket needed, since the three notes in 
the middle all have BeamCount>=2 and look clearly separated:

     3
__________
|  |-|--| |
|  | | -| |
*  * *. * *


[c8 \times 2/3 {c16 c c}] with 1.4's "incorrect" beaming. Looks a bit 
unclear, so we'd better add a bracket:

   |-3-|
_______
| -| |-|
|  | | |
*  * * *

Does this sound good?

Erik