Tuplet bracket position
Alfred's Essential Dictionary of Music Notation recommends that tuplet brackets begin flush left with the first note head (or rest) and end flush right with the last note head. This style recommendation is confirmed by the examples shown in Standard Music Notation Practice (see attached) and the default behavior of both Finale and Sibelius.
MuseScore and LilyPond start and end the bracket at the note stem instead of flush with the note head. However MuseScore's implementation is buggy because this style has a number of exceptions that MuseScore does not cater for (presumably LilyPond does cater for the exceptions but I have not tested it throughly. Adopting Alfred's recommendations would solve the bugs listed below without added complexity to the code because the single rule applies more universally. It also puts MuseScore inline with (what seems to be) the predominate practice.
Bugs that would be solved by adopting Alfred's recommendations:
Comments
The attached "bracket examples.png" comes from Standard music notation practice (PDF)
should work now in rev. 2629
Thanks David.
The following has been made with r2969.
The first inner tuplet bracket is not good. Right?
I missed some special cases. Looks now better in rev. 2633
For beamed tuplets (i.e. no bracket) then the tuplet numeral should be centered on the beam. This is mentioned in Alfred (p.115) but I missed it since it wasn't in the section about brackets. I guess it would have been too easy to have no exceptions :)
For what it is worth Sibelius 6 follows Alfred's recommendations for this special case. My copy of Finale 2007 does not (it matches the current behavior of MuseScore SVN). However, it looks odd to me when the tuplet numeral is not centered over the beam.
I don't have a lot of reference materials with me right now, so feel to fill in or contradict.
The number only tuplet is rendered by simply omitting the brackets. This was ok for the old bracket positions but looks bad now for numbers only. Guess its a new special case :-)
I agree with Comment 6.
fixed in rev. 2660
Automatically closed -- issue fixed for 2 weeks with no activity.