Dear James, Thanks but, again, this is misleading, since the numbers in the bracket would be 6:5 but there would be 3 notes inside the tuplet. I don't have 6 attacks but 3. This is the kind of cheating I do in Sibelius, and that I would like to avoid if possible. Sorry to be so demanding, but isn't this what lilypond should be about? To output exactly what you want? Cheers, Uri On MondayJul 7, at MonJul 723:44 , James E. Bailey wrote: Am 07.07.2008 um 22:14 schrieb Uri Sala:
It is a bit more convoluted than that. I will try to make myself clearer:
I want to write 3 against 5. In this case, since 3 is smaller than 5, it has to use a rhythmic unit twice longer than the unit associated with 5. Let's be overexplicit and call this "3 equally spaced attacks within the duration of 5 sixteenthnotes." Well, I will argue against most people and most notation manuals and most modern scores that the only correct way to notate this is 3 eightnotes against 5 sixteenth notes. Now, neither sibelius nor ENP nor finale allow me to do this, since they force upon the editor the assumed notion that nominator and denominator in a tuplet use the same rhythmic value. This axioma makes it impossible to correctly notate all possible complex tuplets (with nonbinary denominators), and makes the construction of an algorithm that translates proportional notation (a la ENP) into lilypond code incredibly convoluted. I thought lilypond was different  in that I could stipulate different values for nom and denom  but I am not sure now.
Many people would argue that you can use the same value for both, since 3 is so close to 5. But that makes your run into a contradiction. The only way one can stipulate a general and infallible rule for writing tuplets is that a tuplet is the insertion of a certain number of rhythmic values into a space that is smaller. Or, to put it another way, a tuplet  a correct one  is a compression of the duration of a rhythmic value. Very easy to prove: how would you write 6 against 5 sixteenths? Well, just like that. (times 5/6 {c16 c c c c c}. So, if we write 3 against 5, the value that those three notes take should be 8th notes, because all we would have to do is aggregate each 2 sixteenths of the 6:5 into eighth notes. But remember, I want still a total duration of 5 16ths!! So writing times 5/3 {c8 c c} will result in a tuplet twice as long in duration than what I want, since lily thinks that I want the duration to be 5 eighth notes. I have to be able to tell lilypond that I want 3 eighth notes in the space of 5 sixteenth notes (and that is just one of many examples. Trying to to 3 against 7 is even more complicated since the duration is more than twice the attack. In 3 against 7, the three should be notated with quarter notes!). Hope I made myself understood now.
I hope this does not turn into a discussion of the way to notate tuplets, since there is only one that is actually inequivocal and consistent. Unfortunately no editor allows one to produce it, which is a very disturbing fact.
Could lilypond be the one?
cheers
uri
\version "2.11.50" \relative { \time 5/16 c16 d e f g \times 5/6 {a8 f d} }
