Tuplets in 12/8

• Nov 16, 2014 - 18:29

The piece "All I care About" from the musical Chicago is in 12/8 time. There are a few measures where there are "triplets" i.e. three dotted crotchets in the space of two. I know that musically this is equivalent to just writing three ordinary crotchets but I can see the reason for annotating it as "triplets" because that's how it sounds (or how it would sound if it were in 4/4 for instance).

In the attached file, the first measure is how it sounds and the second is how it appears in the original score. I have had to play about with the timings to get it to appear this way. Is there a way to make such triplets properly when in 12/8 time? Would it merit a feature request?

I found a way to do it by creating a dotted minim rest and then selecting it and going to the menu and using Notes -> Tuplets -> Other and making it Relation 3/2 and then entering them as dotted crotchets but shouldn't it create such a Tuplet just by clicking on the dotted minim and pressing [Ctrl] 3 or selecting Notes->Tuplets->Triplet?

Attachment Size
Triplet_12_8.mscz 5.88 KB


I just tried to duplicate this in the recent Nightly rev. c13a719.

Imagine my shock to find that 'Other...' is missing entirely from Notes > Tuplets!

Edit: I just downloaded rev. f783a74, the up-to-the-moment Nightly. Still missing 'Other...'.

I'm concerned that underquark's original post - about what definitely appears to be a bug - not get lost in the shuffle from the tangents we took.

It seems to me that the triplets in question should in fact automatically be rendered as dotted crotchets via the CTRL+3 shortcut. Is there any reason it would be necessary to use the 'Other...' submenu to obtain the intended result when it should predictably happen with CTRL+3 on a dotted minim?

The actual result of CTRL+3 in this case - undotted crotchet triplets - doesn't make sense! If, for the sake of argument, one genuinely wanted to divide the dotted minim into three plain crotchets, they wouldn't actually be tuplets and wouldn't need to be created as tuplets.

In reply to by [DELETED] 448831

I'm not totally convinced it's a bug, FWIW. I agree undotted quarters doesn't make sense in this context - which suggests to me that pressing Ctrl+3 in the first place doesn't make sense. You clearly want something *other* than the normal/usual notation here, so it seems totally natural to me that you'd have to work a little harder for it.

Even if MuseScore were to special case the situation where the division already works out as it does here, it seems to me the natural thing for it to do is to use *half notes* rather than dotted quarters by default.

I mean, sure, we could add special cases both for the unusual situations where you hit Ctrl+3 and the division works out evently and *also* compound that special case with another special case on top of a special case for compound time signatures. But calling it a bug that we don't do this special case on top of a special case for what is an unusual situation to begin with - well, that seems a stretch to me. Minor feature request at best.

In reply to by Marc Sabatella

Well, whatever works - but the workaround of using a relational value is, after all, required to obtain the correct result. Maybe it's worth pointing out in the Handbook that creating tuplets in compound time can't be done according to the method otherwise prescribed.

A documented workaround might obviate the need for even a 'minor' feature request, but I do believe there is absolutely a bug here. MuseScore creates what it thinks are triplets, but - rhythmically and logically - they are not. I can't understand how that is not a bug.

In reply to by [DELETED] 448831

Of course tuplets can be created in compound time. Just not the highly unusual type we are discussing - types on which the basic note value is already divisible by the specified number and hence are completely unnecessary.

To me, the bug is in pressing Ctrl+3 in the first place on a note value that is already perfectly divislbe by three without the need for Ctrl+3. Like I said, you wouoldn't normally need to do that at all, and even if you did, there are three possible things you might have wanted - quarter notes, dotted quarters, or halves. Any of those might be the desired result in some context. MuseScore can't read your mind as to which you actually intended. Best we can do is guess. And like I said, it's a special case on top of a special case.

In reply to by Marc Sabatella

I agree that the situation is unusual.

I disagree that it is completely unnecessary. Whilst it is mathematically correct to express this as three ordinary crotchets (think of "America" in West Side Story), expressing these as a tuplet of three dotted crotchets in this piece does make sense to me in that the purpose of writing music is to convey the way you want it to sound and I think that the tuplets do it better in a 12/8 piece which is counted as if it were 4/4. As it is, when you press [Ctrl] 3 over a minim in 4/4 MuseScore creates three ordinary crotchets and when you do the same thing in 12/8 over a dotted minim it still creates three ordinary crotchets but places a 3 over them and this IS wrong.

I'm not sure I agree with the thinking that MuseScore needs to read your mind here. Normally when you highlight a note (or rest) and press [Ctrl] 3 the tuplet gets expressed as three items each looking like half the value of the parent item. Applying it to a crotchet you get three quavers, applying it to a minim you get three crotchets. Extending this to tuplets in compound time I certainly would expect a dotted minim to yield three dotted crotchets.

There are always going to be unusual situations and keeping the "Other" option in place for tuplets would seem the best way of coping with them.

In reply to by underquark

I didn't mean that this kind of tuplet was "unnecessary" in the sense that it is "unnecessary" we support them. i meant, it is not strictly musically necessary to use a tuplet to express the actual sound. In this sense, it is like a duplet in 6/8 - technically doable as two dotted eighths, but an eighth note duplet is often clearer to convey the "feel". Just as in your cases the triplet might be be clearer to convey the "feel". I do get that.

What I mean by about mind reading is that while in this particular situation, given this particular intent, it might seem "obvious" that the "right" thing to do is use dotted quarters, I'm not that this will be true in all situations for which any "simple" algorithm capable of dealing with this situation might suggest dotted notes.

The current algorithm doesn't special triplets over any other type of tuplet - the calculation always uses simple (non-dotted durations), continually trying smaller ones unless it finds one that makes sense. Dotted notes as the basic unit will never arise from this algorithm.

As a trial, I have tentatively implemented the following:

I let the original algorithm do its thing, but if the caclulated ratio is 3:3 - which is what it is in this case - I override this and make it 3:2, then multiply the basic unit by this amount as well (thus making it a dotted note). So far it seems to work. But I have to say, I'm a little nervous because as I said, dotted note would never have arised as the basic unit from this command before. Since they appear to work when created in Other, though, hopefully it's OK.

If you submit an issue to the tracker, I'll submit the PR. Then we'll leave it to others to see if its OK to merge. And I'd depend on you to really hit this hard in testing, to see if corruption or other problems ever ensue.

In reply to by Marc Sabatella

For the record, Marc - because you said 'you guys' - I wanted to let you know that I was satisfied with the logic of your last reply to my last comment. You said:

"To me, the bug is in pressing Ctrl+3 in the first place on a note value that is already perfectly divislbe by three without the need for Ctrl+3. Like I said, you wouoldn't normally need to do that at all, and even if you did, there are three possible things you might have wanted - quarter notes, dotted quarters, or halves. Any of those might be the desired result in some context."

I hadn't thought this issue through to those other possibilities, and you convinced me that I had jumped the gun in agreeing that there was a 'bug' here.

I apologize for any misunderstanding caused by my not acknowledging my second thoughts immediately.

I have tried entering a duplet of two crotchets in the space of one dotted crotchet, the equivalent of two dotted quavers.

When this is attempted by converting a dotted crotchet or a dotted crotchet rest into a duplet by choosing the Notes > Tuplets > Duplets command (Cmd + 2 on Mac), the result is a duplet consisting of 2 quavers instead of crotchets.

Attempting to use Notes > Tuplets > Other to create a Relation 2/1.5 is not possible as only integral values are accepted.

This was also attempted using Finale Notepad which was able to create the required duplet, however, when this was imported into MuseScore via an xml file, the duplet was written as a quadruplet of 2 crotchets.

This has been attempted using MuseScore 1.3 and MuseScore 2 (beta) with the same result.

Could this be a bug? If not, I'd be happy to be shown a way around this problem.

Thanks in advance.

In reply to by eggwhisk

Hmm. Using Tuplets / Other with use 4 / 3 rather than 2/1.5. would work, except then you would see a "4" instead of a "2" in the bracket. I guess I normally just notate this with two dotted eighths (quavers).

Maybe there is a simple solution I am not thinking of. But for the record, I still have the code I wrote to support the origial request. I never filed a PR because no one submitted a bug report / feature request so I assumed it wasn't considered important. If there is no way to get what you are asking for here, and if my code for that original case also works here - and I suspect it would, or could be made to with a small tweak - then this seems like more incentive to actually make the change.

Do you still have an unanswered question? Please log in first to post your question.