Tuplets continues to be a nightmarish scenaro.

• Apr 21, 2016 - 18:34

You can waste a great deal of time trying to get "tuplets" entered into to Musescore. Often the end result, no matter what you do, is really not what you would want, and is less than ideal. I feel a number of beats needs to be blocked out or "selected" and then the desired number of notes desired be specified to fit in that space. As it is now, it based on "note values." This is okay for basic triplets, but is poor for music such as Chopin wrote where an odd number of notes is forcecd into a a given time duration. Thus the Musescore programmers need to figure out how to base it on beats or portions of beats, not on a specific note value. I continue to feel that the flagging also tends to be wrong and continue to believe that the flags should correspond to the closed even number of notes that would fit in the same space. I certainly pray that this will be fixed in verson 3.0.


Comments

I hate to say it, but have you read the section in the manual on tuplets?

https://musescore.org/en/handbook/tuplet

"To create a tuplet, first select a note on the score that specifies the full duration of the triplet group. (For example, to create a group of triplet eighth notes, first select one quarter note.)
For a triplet, from the main menu, choose Notes → Tuplets → Triplet. This creates a triplet by dividing the full duration into three equal parts."

So, for a 5-tuplet over 2 beats, you put a half note in the piece (choosing the "number of beats") and then choose the number of tuplets you want (in this case 5) by selecting Notes -> Tuplets -> Quintuplet and you'll get 5 notes in the space of 2 beats... for 4 notes in the space of three beats, put in a dotted half note, select it and then choose Notes -> Tuplets -> Quadruplet.

Isn't that kinda exactly the way you said it should be?

In reply to by carneyweb

I don't know, I'll have to try what you are saying. All I know is every time I go about trying to enter some form of tuplets, the end result is not as expected. Also the "flagging" appears wrong to me. For example if you enter 7 notes across on 1/4 note I would expect the flag to be a 32nd., not a 1/16th. As four 1/16ths normally correspond to one 1/4 note; whereas eight 32nds, would be required to replace the time value of a single 1/4 note. (I notce that a septuplet over a 1/4 note gives flags of 16ths, but eight notes in the same space gives 32nds, but a nontuplet in the same space results in 32nd flags. To me the flags should round off to the closest number of even notes in the same space. Obviously less flags is the most desirable option, but only when the number makes visual sense. Seven 16ths following four 16ths doesn't make visual sense in the space of two 1/4 notes, as the group of seven is played almost twice as fast at the first group of four. Is the rule to always round down, not up? Could the end user at least have the option to specify the value of the flag in a given group?

(As an aside, I'm traveling today and won't be able to continue in the discussion any more today.

In reply to by gBouchard

If after reading the Handbook regarding tuplets, and considering the comments given here, you are still having trouble, then feel free to post the specific score you are having problems with and descirbe precisely what you are trying to do, and someone should be able to explain it to you.

In general, MuseScore uses pretty standard ways of notating tuplets. 7 notes in the space of 1 beat is normally notated as 16ths, not 32nds. The rule, as mentioned, is to notate using the next *longer* value, not the next *shorter* value. The case of triplets is most obvious - if music followed your rule, then three notes in the space of one beat would be notated as a sixteenth, but we all know that isn't how it is done. Three notes in the space of one beat are notated as eighths. 5-7 notes in the space of one beat are notated as 16th, 9-15 as 32nds, etc.

However, it is indeed true that some editors prefer notating septuplets in particular using the next smaller note value. You can force MuseScore to use this convention by entering the tuplet manually - Notes / Tuplets / Other, enter 7:8 as the ratio.

In reply to by Marc Sabatella

FWIW, if anyone is looking for an "official" reference on the subject of how to notate tuplets, Elaine Gould's "Behind Bars" discusses this on p. 203-4. She presents both options - the one where you always use the next larger value (so seven notes in the space of one quarter note are notated with sixteenths) but also the one where you "round" the values (so seven notes in the space of one quarter are notated with thirty-seconds). She gives reasons why the first option should be preferred ("recommended as the clearest notation"), but the implication is that second option, while less desirable, is still considered valid.

This is the rule MsueScore follows. By default, tuplets are notated the recommended way, but the less desirable but still technically valid method can be achieved using the dialog.

As said, it would be very helpful to have one of the Chopin sheet you are referring to, or a particular measure you are struggling with.

In reply to by ChurchOrganist

I did watch it, and find it confusing regarding the Chopin portion. Also, and don't mean this as an insult, but as constructive criticism, all the organ music at the beginning and end wastes too much time for a tutorial explanation. Just play the three note motive. It's enough to suggest you are a Church Organist. As for 3 eights in the space of 1/4, in my opinion 3 eights is as close to 2 eights as it is to 4 sixteenths. I said in some other thread that the obvious best rule would be to use a single flag bar as opposed to a two bar flag. I presume the general rule is to choose the lesser number of flag bars when the note count falls between two possible choices. However I find Musescore to be inconsistent itself or at least it's always very unpredictable to me and very hard to anticipate what is going to be generated and the look of the music in the end can be rather poor as well. I honestly feel if it were based on beats as specified by the time signature it would be far more comprehensible. Finally with the inspector perhaps the end user could be given the option to select the group and choose for himself the number of flag bars so that he could unify the look of the composition. As it is now, I find myself experimenting over and over again with many various results and most of which are rather miserable in final result.

In reply to by gBouchard

See my reply earlier in https://musescore.org/en/node/107576#comment-485046. The rule is quite clearly stated in Gould, and we follow it correctly and consistently as far as I can tell. The "rounding off" method you suggest is definitely not as recommended, which is why it is not the default in MuseScore, but as I described, you can easily create that if you like by simply entering the ratio the way you prefer.

Not sure what you mean by unpredictable. As far as I can tell, it always does exactly what you tell it. I suspect it's just a case of you not being familair with it yet so are doing things more by trial and error. But if you find a situation where it does not work the way it is designed to - which is to say, consistently following the rule from Gould - feell free to give us precise steps to reproduce the problem.

Also not sure what you mean about being based on beats as specified by the time signature - many things in MsueScore *are* based on this. Again, if you have a specific situation where things are not, please describe in more detail.

In reply to by gBouchard

So in particular, if you want seven notes in a beat to display with three beams - as if they were thirty-seconds - that's saying 7 notes in the space of 8 (since normally there are 8 thirty-seconds in a beat). Thus, you enter a ratio of 7:8. Whereas the normal shortcut for creating septuplets will use the more standard interpretation of 7:4 (7 notes in the space of 4 - which is to say, notated as if they were sixteenths, which normally come 4 to a beat).

The ratio notation is standard, BTW, it's not some obscure MuseScore UI thing. Lots of published music will print the full ratio in the tuplet, for complex ones anyhow. That's why this is also given as an option in the dialog.

In reply to by Marc Sabatella

Okay thanks, I'll try that. Maybe my understanding of the concept will improve as well as the final results of the Musescore documents I'm trying to create, because I was having trouble getting an appropriate flag count (visually appealing or appropriate) for the various tuplet figures I was creating.

I am having a lot of difficulty trying to create a score for a piece by Brian Ferneyhough. I agree that there seems to be an easy solution: allow a tuplet to be created over a prescribed amount of musical time, instead of just standardized musical note-values. For example, in measure 54 of his String Quartet No. 6, the second violin, viola, and cello all have the same rhythm. The time-signature is 9/32, and the passage in question is a nested tuplet of 8:5 32nd-notes under 6:4 32nd-notes. It was easy to create the 6:4, but creating the 8:5 is impossible because there is no single note which equals 5 32nds. It would be great if a user could tie together whatever combination of notes creates the proper total length, and then use _that_ to make the tuplet.

In reply to by monz

So I stated my problem there, with a suggestion for an easy solution. But is there _any_ way to do this, even if it is difficult? How can one make a tuplet to fit into a note-value that is not a power-of-2 or a related dotted value?

In reply to by monz

Create an extra dummy measure, set its length to 10/32 (or whatever), create the tuplet on the measure rest (you may need to select the measure and hit Ctrl+Shift+Delete to convert the normal rests into a measure rest), then copy it to where you want.

In reply to by Marc Sabatella

Thanks, Marc, your reply is one which I suppose would work for a single x:5 tuplet. But in this particular case the 8:5 tuplet is nested _under_ the larger 6:4 tuplet ... in other words, the entire measure has 9 32nd-notes, of which 4 are changed into a 6-tuplet; then 5 _of those_ are changed into an 8-tuplet. Would your method work in this case? Just thought I'd ask before I take the time to try it.

In reply to by monz

I believe that from an intuitive point of view it would be better if Musescore had the end user select the number of beats or portions of beats as a method for entering Tuplets. I can understand now why it is the way it is, and I believe that's because it's the way programmers think about it. But I think performers generally think about a number of notes within a beat.

In reply to by gBouchard

One thing I'd like in this context is an easy way to set up multiple tuplets: If two measures of triplets are coming up I'd like to select the two measures and turn all crotchets in them into triplets with one set of commands. Then I'd go ahead and fill in the pitches. I do that now and set up a series of empty triplets before dealing with the pitches, but it is slow work.

In reply to by azumbrunn

You can already do this. Place crotchet rests in the measures. Select the crotchets and press [Ctrl]3. Everything you select will, in fact, be "tripletised" - crotchets will become 3 triplet quavers, minims will become 3 triplet crotchets. Select some dotted crotchets in a 6/8 piece, press [Ctrl]2 and they will become doublets etc.

Do you meant you'd like to create triplet quavers and have the duration stay that way until a different duration is selected? That has been requested in the past but not implemented as far as I know.

In reply to by gBouchard

Realistically, in the vast majority of cases, there is no difference between selecting "number of beats" and selecting "total duration". It's exactly the same thing except in the very cases where the total number of beats is not representable by a single note duration, and then suddenly it becomes impossible. Since we already have a clear way of selecting note duratiosn that people already have learned (using the toolbar, or the keybaord shortcuts, or custom MIDI triggers) it makes sense to use those so users don't need to learn a whole new interface just to create basic tuplets. But perhaps the "Other" dialog - currently only needed for especially complex tuplets - could be extended to allow the number of beats to be specified in the same way that we do for pickup measures et al. That might be a relatively easier to implement and yet reasonably intuitive way around this rare problem.

In reply to by Marc Sabatella

I'm of the opinion that a majority of Musescore users are often frustrated by its unexpected behavior. I notice that I'm often required to re-enter various notes a second or third time because Musescore makes unexpected changes to notes that were once originally correct and not in need of any change whatsoever. In short Musescore is often overly automated. Tuplets seems to be a prime example, but it certainly is not the only area of concern. i am cognisant that the programming issues are immense. I'm not confident the program will ever be considered actually user friendly. The mathematics of all the manipulations seems to be overwhelming.

In reply to by gBouchard

If you are seeing unexpected changes, then eityher this is a bug you need to report, or else something you are not understanding correctly about what the expected result of your action should be. Either way, we can't do anything to help without a sample score and precise steps to reproduce the problem you are perceiving.

Most users never enter any tuplets other than simple eighth or quarter note triplets, and as far as I know there are no known issues with these.

In reply to by gBouchard

I don't understand the difference. We already have a command to specify "half note" as the duration (for example) - an icon on the toolbar and a corresponding keybaord shortcut. Are you suggesitng we not simply allow users to use that existing command they have already learned, and instead force them to learn some *new* command - and provide a new GUI element and/or a new keybaord command - in order to specify the duration of tuplet? This seems a step *backwards*. Why have two *different* ways of specifying a duration, depending on whether you plan to enter a single note or a tuplet? That seems unnecessarily complex.

In reply to by gBouchard

I don't understand the difference. We already have a command to specify "half note" as the duration (for example) - an icon on the toolbar and a corresponding keyboard shortcut. Are you suggesting we stop simply allow users to use that existing command they have already learned, and instead force them to learn some *new* command - and provide a new GUI element (toolbar icon or dialog box) and/or a new keyboard command - in order to specify the duration of the tuplet? This seems a step *backwards*. Why force people to deal with two *different* ways of specifying a duration, depending on whether you plan to enter a single note or a tuplet? That seems unnecessarily complex, except as mentioned in the rare case where the duration cannot be specified as a single note value. But why penalize all users trying to enter simple triplets just because in some rare cases the simple existing method doesn't work?

In reply to by Marc Sabatella

I'm not for penalizing anyone. A half note as I understand it relates to the time Signature. If it's in 4/4 it gets two beats yet in 2/2 or "Cut time." it gets one beat. The problem with complex triplets is they are not a just a single note, but a group of notes played over a defined time duration. It just makes more sense to base them on beats rather than notes, as you have to have the correct note type selected in Musescore before you can choose the triplet, but the logical thought process is to think: "I want a certain number of notes within this time period meaning number of beats. Not I want as certain number of notes to replace this single note. Obviously they are interdependent, but the fact is the way it works now is not particularly intuitive to the average person.

This argument reminds of when I first started to use Linux and much of what needed to be done had to be done with the command line for reasons of necessity. The Linux "experts" at the time were adamant that the command line was the way do go, as a graphical interface was to them unnecessary and wasteful.

In the end the GUI won out. I'm quite confident that in some future version of Musescore, triplets will be entered based on a specified number of beats as it applies to the time signature, not a note type that itself changes in rhythmic value when the time signature changes. In other words half notes normally have a value of either one or two beats based on the time signature. But beats are beats no matter what time signature is specified.

It's really so basic; and I'm sure you well know this as you are a teacher of music theory at the University level, as well as a very fine piano player, musical composer, and technical writer.

There is a a book by Hal Galper (http://www.halgalper.com/forward-motion/) that suggests that jazz phrasing is actually based on a feeling to two as opposed to four, which is perhaps is the more common view.

I guess it sort of goes to snapping your fingers on 2 and 4, not 1 and 3 which could land you in the penalty box with "hip" players.

In reply to by gBouchard

Let me as just another average musician chime in now. To me, basing tuplets on total note value is far more logical than on 'beats'. Although, as you point out, those are related anyway.

Your approach might work for some common time signatures where the beat always had the same length (like 4/4 or 6/8), but it seems to fail as soon as not each 'beat' within a measure is of equal length (take 7/8 or 5/8) or having compound meters.

The current method works in all of those situations in the exact same way.

To me, the only part missing is that creating a tuplet would work on a tied note and then uses the total value of the tied note for those rare cases you'd need 7 notes over 5 eigths duration for example.

In reply to by gBouchard

Either I don't understand the distinction you are making, or I disagree with it. I think cut time actually makes a marevlous example for why the current system is superior. You want to have to enter quarter note triplets differently in 2/2 than in 4/4? That seems crazy to me. Quarter notes are entered the same in either time signature (type 5 Ctrl+3, or click quarter note icon followed by Notes / Tuplets / Triplet). Why should it be different for a quarter note triplet? A sequence of clicks / keystrokes that successfully creates a quarter note triplet in 4/4 should work exactly the same in 2/2 or I think the average user - even the advanced user - would be extremely displeased.

So actually, I think your example is about the clearest arguyment *against* your proposal I can think of.

In reply to by Marc Sabatella

It's not that the present system doesn't work, it does. But it requires special tutoring as the help you gave me at some point in understanding how to enter the tuplets you actually want. It's not intuitive. Obviously their are bigger brains than mine working on Musescore, and yours is most likely one of them. However the present user interface is difficult to comprehend without detailed explanation. Apparently even your book does not go into the subject in adequate detail. All I know is that there often is a lot of frustration trying to do what should be the simplest thing in Musescore. It's not that it can't be done, it's more that it's very hard to figure out to do it. As to whether cut time is exactly the same as 4/4 notation wise, I cannot say. The guy who points out that complex or compound meters like 6/8 essentially are based on two beats rather than six, makes a good point.

Regarding Bach's Goldberg Variations, the version I have which came from the Musescore Website has a time signature of 3/4, but I play Variation 25 with a feeling of 9/8 where an 1/8 note gets one beat or count. So yes there is a clear difference between beats and note values when it comes to actual performance. The problem with Musescore is the user interface is often user unfriendly is certain areas. You have said there are major changes coming in Version 3. I hope that I find it easier to accomplish simple tasks. Or at least it is more intuitive without the need for special tutorials.

In reply to by gBouchard

One more thing: some things are complicated. No way to make them simple. There is only one way to master them: Take the time to learn (as you say).
I am somewhat allergic to the term "intuitive" as meaning "easy to use". It is Apple's shtick to call its user surface "intuitive" when in fact it is often bizarre to minds other than Steve Jobs's (check out iTunes!).

In reply to by Marc Sabatella

Hello Marc,
before stating my opinion, I want to thank the creators of MuseScore in general and greatly appreciate providing an amazing tool and a great free software.
Regarding Tuplets: The given options, only creating Tuplets on notes having a duration as a power of 2 (plus dotted notes) is really limited. Coming from a powertab/guitarpro background, where you could choose both numerator and denominator freely (e.g. creating 8:5 tuplets and nested tuplets with ease), it is really a frustating experience working with musescore on that. While the shortcut of selecting a note/break with the duration which the tuplet should be in is fast, it wasnt that intuitive to me. Being used to deal with all kinds of tuplets, I really want the freedom to choose the denominator as I wish and not being stuck to first create a corresponding note/rest and second can choose any amount of notes in any duration I want. There is no way to create my 7:5 tuplets, 15:9, 13:11 and so on so forth. This is just straight up a missing feature and should be implemented as soon as possible, no matter how many clicks it takes to write these. People, musicians, really need to be able to freely write down any kind of tuplets.

In reply to by Ryan Stark

"There is no way to create my 7:5 tuplets, 15:9, 13:11 and so on so forth"

This makes no sense to me. What is a 15:9 tuplet for example. What is the total duration of the tuplet? What duration does the 9 refer to? I'm asking because I know of no limitations on tuplets in MuseScore. It might not be as fast as other programs to make the tuplet, but all are possible. There may be some exceptions in nested tuplets, but I'm not positive.

In reply to by mike320

If you want to put 15 notes in the same space as 9 notes in any meter, that is 15:9. If you have a 4/4 and you want to play 15 notes in the space of the first 9 16ths, that will be a 15:9. In the following thing, 15 notes over the area from 1 up to and including 3:

15 notes here
|........................| (I hope it shows right in your browser)
1 e + a 2 e + a 3 e + a 4 e + a

This is an easy thing, as it is three consecutive 5:3 in this case. Since 15:9 is three 5:3, you can get away with it in this case, as 3 16th notes form a dotted 8th note, and it is doable. However, if you want to do something else, like 17:9, there is no way to do this, because there is no way to write 9 notes as one rest. The usual problem.

In reply to by sami.amiris

The way I do this is I use a measure after the last notes entered in my score and make a measure with the duration I need by right clicking the measure, select measure properties, change actual duration to what I need, 9/16 in your example, and press OK. You now have a whole measure rest with an actual duration of 9/16. Select that rest and use note->tuplet... to set up your relationship. Copy the measure to where you want it, then select the measure and press ctrl-delete so it doesn't mess you up later. Like I said, not instantaneous but not impossible either.

In reply to by sami.amiris

I just created a "How to" with an easier to explain method for those with less understanding of how to use MuseScore. These are advanced techniques and are not standard in most music, so they do not belong in the handbook, but a "How to" or "Tutorial" is a good spot for the explanation. Also, everything you need to be able to do this is in the handbook - just not on the same page. I figured most of these methods out on my own.

https://musescore.org/en/node/260646

In reply to by Ryan Stark

I would certainly agree that having an easier way to create complex / irregular tuplets would be valuable to some subset of users. As with all features, the trick is in figuring out how to prioritize it with the other features other users are asking for. Different people "really need" different things, and what one persona finds missing and wants to see implemented "as soon as possible" is not always how someone else sees it. The more often it comes up in the forums, the higher it will likely rise in priority.

BTW, creating nested tuplets is already extremely easy, not sure hwo that could be reasonably improved. But triplets where the total value is not something that can be represented by a single note / rest value is indeed not currently very convenient. Probably 99% of users never would use such a feature, but I can easily believe the other 1% would use it a lot.

In reply to by gBouchard

If Musescore alters notes already entered in response to a follow up entry it must be that the software perceives the new command to imply a correction to earlier entries. I suspect that whenever that happens to you it is in essentially the same situation. So next time it happens why don't you save the score (or the piece of it where the error occurs) and also make sure you describe precisely the sequence of entries you made and post it for the experts to look at (the score maybe before and after the error)? I am certain the problem can be solved.

IMO the program is already user friendly, there is just some misunderstanding going on somewhere--happens to the best people sometimes.

Actually it is not a rare problem at all. A huge portion of post-19th-century piano, chamber or orchestral music is literally infested with polyrhythms. And many of these are either stuff like 6:5 (Scriabin), 11:6 (Chopin!), 7:5 etc., or nested polyrhythms, which exists in virtually HUNDREDS of pieces for piano (for example, Stockhausen's piano piece No 2, first measure, and on). With Xenakis, Ferneyhough etc., even worse. And they are not rare, they are pretty common in the circles of contemporary composition today. Not for the faint of heart (it takes quite a few years of intensive training to be able to understand, hear and play these correctly) but most certainly doable, and actually very much done today.

I need all of these in my music:

i) x:5. This is simply a must. Also, in the same vein: x:11 etc. There is no single value as long as 5 or 11 notes, so entering such ratios seems hard in comparison to the competition out there (including Lilypond which is free!). For example, meter is 15/8 = 3 x 5/8, and I need three (quarter) notes over the last 5/8 in the measure.
ii) nested polyrhythms, for example 16 eighth-notes over 11 eighth-notes which themselves are part of 15 eighth-notes over 8).

Yes, I actually perform such patterns regularly, and I am not anything special, there are many, many people doing this out there. Stuff like that does NOT equal so-called 'paper music'. I need them because I use them and play them.

So: How can we notate these correctly?

P.S.: Sometimes It is not enough to contain just a ratio, because that simply takes a long time to do a computation of the overall duration. In these cases it is simply easier for the performer and clearer to everyone involved to include actual small-print values in the ratio as well, for example 13(eighth notes) : 5 (quarter notes), if the rhythmic patterns used are overly complex.

P.S.2: Sometimes a smaller number on the numerator is exactly what you need. For example, if you have a meter of 10/8 = 2 x 5/8 and play three equal notes over the last 5/8, it is convenient and correct to write 3 (quarter notes) : 5 (eighth notes). The quarter notes are in actuality 1/3 x 5/8 = 5/24 each, a tad smaller than a 'real' quarter note, since 1/4 = 1x6 / 4x6 = 6/24, equal in length to 6 16th-note sextuplets, whereas our 3:5 quarter is equal to the first five of the said group.

In reply to by sami.amiris

It seems your basic problem is creating ratios in MuseScore.

I'm going to be very simplistic with the explanation, so don't be offended if it is too simplistic.

Under the notes menu is a selection for tuplets then other.... Click that.

This will bring up a dialog that allows you to set the tuplet to what ever ratio you like. You need to determine what the duration of the overall tuplet is. It must be a value that can be expressed in a single note. So making a tuplet out of 5 16th notes is not impossible, just more difficult. Once you determine the overall duration of the tuplet you can enter the ration you want.

You can then decide how you want the tuplet numbers and brackets to be displayed. You can select number, ratio or none. The bracket can also be turn on or off or let MuseScore decide for you. You can change any of these option in the inspector later if you want to.

For nested tuplets, the process is the same. Once you make your main tuplet, select a duration and set the tuplet inside the existing tuplet the same way. You must work from the outside to the inside.

If you want straight forward tuplets from 2 to 9, pressing ctrl-# (#=number of tuplets) to create it.

Assigning a duration has 2 options. while in note input mode you can select the duration you want by pressing it. Suppose you want a dotted 1/4 note duration with a 4th-let in it. Press:

5 . ctrl-4

This will create the desired tuplet.

If note in note input mode, you can select a rest of the proper duration and then create the tuplet. If you have a whole rest and want a 1/2 note long tuplet, select the whole rest, press 6 then assign the tuplet to it.

Both of the above methods work for nested tuplets.

Working with existing tuplets:

There are issues with tuplets that you need to be aware of.
1. you cannot copy or paste part of a tuplet.
2. you cannot select any part of or and entire tuplet and paste anything over it. If you are pasting an entire measure and the destination has any tuplet in it, you must delete the contents of the measure before pasting. You also can't use swap with clipboard in this situation.
3. For tuplets that are ratios, copying and pasting works fine...until you save your file. At that point, many tuplets become impossible to copy or delete without deleting the entire measure first. This has to do with the way tuplets are internally calculated. Hopefully this will be fixed in the future.
4. Not all is bad, you can change durations of beats within a tuplet like all other notes. You just can't make a note/rest extend past the end of the tuplet. You need to enter a note and use a tie.
5. Commonly encountered tuplets copy and paste with no issues.

If there is anything that you encounter that is not covered here, feel free to comment. Someone will help you.

In reply to by mike320

How could I ever be offended by your good will and honest effort to do something constructive? Absolutely no offense. And your answer is not simplistic at all btw. It is fine.

Could you be more specific on the 5 16ths issue? For a real world example, let's make it 5 eighth notes:

Meter: 10/8 = 2 x 5/8.

Tuplet = 3 (quarter notes) : 5 (the second set of 5 eighth notes in the meter above).

It would look like this:

_______________1____2____3____ - The 3 equal notes I want to notate (3:5), say notated in quarter notes.
1__2__3__4__5__1__2__3__4__5__ - The 10/8 meter, counted as two sets of normal 5/8 (not 8th-note 5-plets!)
1______________2______________ - the 5/8 groups. The 3 notes I want are on the second 5/8 group.

The problem: I cannot find a rest that equals the last 5/8, so my tuplet will not have the correct overall length.

I can put three quarter notes in a HALF note and then play around with the brackets to fit the measure and all, but the ratio is not really 3:5 in this case - it is actually 3:4, as there are 4 eighth notes in the half-note.

If I push my luck writing 3:5, it would expect to find 5 notes in the half note. Which notes are 5 in a half note?

Eighth-note quintuplets! Not normal eighth notes.

So the overall length of my tuplet would be:

*) 3 normal triplet quarter notes (over a set of quintuplets, which makes absolutely no difference here, since they equal a half note anyway!),

and not

**) 3 notes in the space of 5 normal 8th notes, which equal a half-note plus an eighth note.

The latter are a bit slower than the former. Each of the former is 1/6 in length, each of the latter is 5/24. To make the comparisons, 1/1 is the whole note, and 1/4 = 6/24 is the normal quarter note.

Am I right to assume the behavior of the program, based on the 'one rest per tuplet' mandate?

If yes, is there a workaround with this?

If yes:

i) can the steps be outlined so that I can replicate them?

and

ii) is the outcome of the workaround correctly playable by the program?

Thank you!

In reply to by mike320

Hi, and thank you for your very helpful answer!

in essence, you seem to break the measure in smaller submeasures and since the whole measure, no matter the length, has one empty rest, you define any tuplet you want in that! Then you recombine the smaller submeasures into a measure of the original length and you are done.

Pretty ingenious!!

It should be part of the manual, btw!

One question:

Can this method be generalized for nested polyrhythms? Is there an example of this nature?

For example: 4:5 and the 5 is a part of a 7:11? The numbers are chosen so that there is no easy way for one rest to cover them.

The base 7:11 can be done with your method. How can we select the first 5 of the 7 to do the 4:5 over them?

Thanks!

-S.

In reply to by sami.amiris

I don't believe nested tuplets like this is possible since it would require having tuplets cross measure lines - which is impossible anywhere (musescore or actual music notation). The alternative would be to paste something into a tuplet, but this is not allowed by MuseScore.

I have never seen a professional score that makes a tuplet out of a duration that cannot be expressed by a single symbol. You say there are hundreds. Could you post some public domain links to a couple of them so I can examine them?

As far as the handbook is concerned, this is an advanced technique and should be put into a how to page that is linked to on the tuplet handbook page. Currently, How To's are almost impossible to find, so I'm not going to waste my time with a how to that no one will use. Once the MuseScore.org site is upgraded in the next couple of months, I will create a how to for this and a few other advanced work-a-rounds that have often come up in the forums. The upgraded site is supposed to make it easier to find pages that are now buried deep in the abyss.

In reply to by mike320

Public domain means pretty old. Especially with copyright extensions and all that. And these techniques are relatively new, so I cannot give you any PD links easily. I don't know if even Roslavets is PD in the US.

Examples of pieces (among many):

Stockhausen: Piano piece Nr. 1, the whole thing (there is a video of it on youtube with the score and sound). Also No 2, etc.
The Darmstadt school is full of stuff like that.
Zappa: The Black Page, many others.
Luigi Nono: Sofferte Onde Serene, many others
Xenakis: many, many pieces.
Fenneyhough: way many by him, and of people in that general direction, like Michael Finissi etc.

etc. etc.

All of the above decisively non-PD.

Works by contemporary composers are literally infested with such things. Of course, mostly non-pd.

As a postscript, I think that more than a few people will (at least incrementally) search for this. At least from the modern jazz and contemporary classical fields. And they are some of the most enthusiastic supporters of MuseScore, exactly because of its free and open-source nature, and ease of use.

Nobody expects a new L!lyp0nd (which can do almost anything imaginable), but with just a few additions it could come pretty close, AND be easy to use too. Can't get any better!

In reply to by sami.amiris

I believe I understand the concern. When I said I this was a rare occurrence, I guess I meant in the broad context of the world of music at large, but it is true that if you focus primarily on 20th / 21st century classical music, you'll see it disproportionately often. And it's still the case that for the "normal" case, the current method works well.

Anyhow, I agree it would nice nice to have a more efficient way of dealing with these cases where the total length cannot be specified using the duration toolbar. Of the various ideas that have come up regarding specific improvements, do you have any favorites, or new ideas?

Here's what seems most natural to me:

Currently, the Tuplets / Other dialog continues to use the toolbar to set the total length, and there is no override. What if the dialog also included spin boxes to set a fraction (akin to how one sets actual duration in Measure Properties)? The spin boxes would default to the duration specified on the toolbar so the usual case would involve no change to the use model at all, but a twerak to a spin box would give you the result you wanted for the irregular cases.

Something like that make sense? I suspect this would be quite easy to implement, although we're still struggling with #202271: Copy-paste sextuplets/octuplets and their removal leads to corruption which involves some of the same code, and it probably doesn't make sense to look at this until we sort out our strategy for that.

In reply to by Marc Sabatella

I think something in the tuplets dialog that allows you to set the overall duration of the tuplet would be a very good idea. The method used in the measure properties dialog work wells IMO. It would of course need to be smart so you can't set the duration to extend past the next barline.

I would suggest that the current method of dealing with tuplets in general is flawed and should be revamped, otherwise there would not be so many problems dealing with tuplets. One example is local time signatures, which from a users point of view appears to use the tuplet code.

In reply to by mike320

Actually, holding the polyrhythms inside the barline is not really a requirement. This is solidly evidenced by Scriabin's Op11 No1, the very first two pages of this:

http://hz.imslp.info/files/imglnks/usimg/6/62/IMSLP10496-Scriabin_-_Op…

in the second page, last 3 systems, the quarter-note triplets are shifted by two 8th-note quintuplets to the left, and they are rendered over the barline repeatedly. This of course means that the first note of the quarter-note triplets is synchronized with the 9th note of the eighth-note quintuplets, which means that the next quarter-note of the triplet, the one that shows to be on the downbeat, is not really there, but comes slightly earlier, inside the previous measure, because of the 3:5 polyrhythm in effect.

The same situation arises in quite a few others, such as the Xenakis example I mentioned above, measure #10.

Imho, that should be a question: "This rhythm exceeds the barline. Do you want to:

-keep it that way or
-cut it inside the measure, or
-keep it but in a more 'metric' written form?

And press the corresponding buttons to get what you want. Scriabin and Xenakis would have chosen the first option...

In reply to by sami.amiris

In Scriabin's prelude, given the irregular grouping of beams and lack of tuplet marks it is rather subjective as to the start and end of each tuplet throughout the entire piece. There are consistently 10 1/4 notes throughout the piece rather than the expected 8 prior to the last 3 systems. All of the measures can be notated using a standard tuplet for the entire measure. The rubato and accelerando sections make it difficult to determine if there were irregular tuplets used in writing the piece.

Xenakis clearly does extend tuplets over barlines. I have never seen this before. This is apparently part of the evolution of musical notation that is not considered standard. What is extremely unusual in measure 20 is that the duration of the 2nd 1/8 note in the final 3:2 tuplet definitely crosses the barline.

Irregular lines of music with multiple tempos being played at the same time have been written since the middle of the 20th century, Aaron Copeland "The Curlew" is the first example I can think of. The notations of these have varied from composer to composer much as classical notation started with a variety of notation systems prior to their being standardized.

I think that notating the unusual to represent modern music should be a goal of MuseScore in the future.

In reply to by mike320

For the Scriabin's piece the time signature is 2/2, i.e. one whole note per measure. The right hand plays ten notes per measure, so it is 8th-note quintuplets. The piece starts at the fourth note of the second quintuplet of the pick-up measure. This is to give the listener the false idea that the piece has a downbeat there than the actual downbeat, one of the first time illusions in history - like the ones very common in jazz nowadays, but a century earlier.

The quarter-note triplets of the left hand (at the third system from the bottom and on) start exactly at that very place, and for the very same reason. This means that they start every fourth of the five notes in the quintuplet, and that is the note they synchronize on. The phenomenon is not unique to Scriabin, many have used it , but in Scriabin's time it must have been one of the first instances of the phenomenon. In this paper:

http://www.johnlinkmusic.com/JohnLinkDiss.pdf

John Link calls them "out-of-phase" polyrhythms. They are used quite a lot in Elliot Carter's music.

All this discussion is to prove that the triplets are in fact over the barline, and correctly notated so.

All this shifting has a much longer history, for example in Schumann and Brahms among others (a great book entitled "Fantasy Pieces" by Harald Krebs deals with precisely this very phenomenon). However, with quintuplets, Scriabin must be one of the very first to have done this, if not the actual first.

With the Xenakis example:

https://www.youtube.com/watch?v=ZH4j70KU-RQ

at measure 10, the tenor voice is - I think, due to the ratio - out of sync with the others. In any case, measures 16-17 are most definitely infested with polyrhythms that extend over the barline.

In reply to by mike320

I agree. Being able to meet the needs of modern music will give MuseScore an edge inside the Academic institutions, and that can be nothing bad. There us always fresh blood that will take to the program. The open-source and free nature are irresistible for any student of composition that also has to make ends meet. The seeds will be there for the people to continue with the program later on. Plus, the songbook on the tablet make a great duo...

In reply to by Marc Sabatella

Hello, and thank you for your reply.

It is a great fact that the program allows time signatures such as 3/5 - please don't disable this, it is a most welcome feature! Which means that the machinery to write any and all polyrhythms, nested or not, complete or incomplete, inside meters with binary or non-binary denominator, is actually already here in some form, as it seems from a first glance. Usually what is at fault is an attempt to 'simplify' things, which always ends up making them more complicated.

In effect, I believe that the answer should as all-comprehensive and simple as possible. That could be many things.

1) One could specify:
-a point of start and a point of end of the region that the polyrhythm is to be put in, or
-a starting point and a precisely annotated duration, such as 3/5 or 4/6 etc. - always in reference to the whole note as 1/1) or
-have both methods(!)

and put as many notes (which should be user definable or at least selectable) in there as desired. The same thing is - in theory at least -
used already in meter such as 3/5 etc. That would TRULY be awesome! And for really hard-to-write rhythms, extremely practical.

2) Also, one could select a number of notes or rests already in existence in the measure and apply a given ratio over them. This could then work recursively for nested polyrhythms, ad infinitum. That would also be awesome, and practical.

3) One could actually be able to select any of the above methods and that would truly be the best.
As long as meter and ratios are treated in the same 'open' manner, it would truly be great.

Now, most casual users wouldn't really care about all these, so the easiest solution would be to hide all these inside an 'advanced' tab in the tuplets section, and be done with it.

A couple of pointers:

1) The polyrhythms may or may not be bound by the barline. I have given examples of it in this same thread. As there are groupings that are more easily comprehensible if they are written over the barline, the same is true of polyrhythms, and of course polyrhythmic groupings or grouped polyrhythms. It would normally be a wise option to allow for the groupings and/or polyrhythms to be written over the barline or not, according to the wishes of the composer.

2) There should not exist any type of barrier in the number of recursions one uses in nested polyrthyhms. That should be left to the composer. If it ends up being only computer-playable, that is the composers problem, or wish.

3) If the programming strategies used to create the more advanced material are in conflict with the easier general practice, there is always the solution of "hide it in an advanced tab" and keep it simple for those that don't need it. Which means that, even if the programming underneath all tuplets change, getting triplets should be as easy as 1-2-3.

4) The program should not choose the value for the polyrhythm, but should leave that choice to the composer, at least within some boundaries. For example, on the 3:5 that I managed to create with mike320's help, the 3 value is an 8th note, but I would want it to be a quarter note, as it is so much closer to that than the 8th note it wrote.

5) It would be a great idea, as in the Xenakis example in the videos I posted, to include one more mode in the ratios: x value_1 : y value_2, where value_1 and value_2 would be quarter notes, eighth notes etc. So instead of a simple 3:5 I could write 3 Q : 5 E (replace quarter note for Q and 8th note for E). This would make things so much clearer for the performer. Again, as an option.

A sidenote: Some people think that all these make the scores so complicated that the performers end up confused and not knowing how to play the music. In reality it is quite the opposite. The music IS complex for sure. If a performer has spent the appropriate number of years needed to master such music, these things actually are immensely helpful, as they clarify what is going on. They will not help the uninitiated, but they will help the professional performer a great deal. At least as far as I have seen.

If I think of anything else, I will say it.

Tx!

-S.

In reply to by sami.amiris

A general comment:

Changing the overall architecture of how tuplets work is *possible* but beyond the scope of anything I would be prepared to deal with myself. I have a sense this is something Werner has been thinking about and maybe some groundwork is being laid for 3.0, but I can't speak to that. So allowing tuplets across bar lines, while a useful thing I would agree - is not likely to happen by my hand. You can already fake this by creating a long measure then inserting or moving the barline graphically, so as a workaround this will have to do for now.

But a way to simplify the process of creating tuplets with a duration that can't be expressed using the duration icons on the toolbar is something that can be implemented with no big architectural changes, once we agree on the most logical UI for this.

The idea of allowing you to select a region as a way of expressing the total duration is not without merit. However, I see two drawbacks. One is that it requires to have already entered "dummy" notes or rests to cover that exact duration. The other is that we already have a very useful behavior if you invoke a tuplet command with a range selected: the tuplet gets applied individually to each note or rest element in the selection. This is a fantastic way to create a whole row of triplets, for example, and we can't inconvenience that extremely common case just to simplify this much less common one.

On the other hand, while the "regular" tuplet commands currently have this useful feature of working on selections, Notes / Tuplets / Other does not - it requires a single note to be selected. So actually, just making that dialog operate on a selection as a way of setting the total duration is not out of the question. But then the question is, what would become of the notes already in that selection? Easiest would be to simply clear then out.

Which brings me back to my previous suggestion of allowing you to specify a total duration in the dialog. This actually would have the same issue - now that it is not operating on a single selected value, it might cross into existing notes or rests, and we have to decide what to do with them. So, imagine, with a half note selected, you try replacing the first 1.25 beats with a tuplet. Again, easiest is if we simply deleted the half note. If you want a dotted eighth after your tuplet, you can enter it again afterwards.

As for choosing the value used to represent the notes: this is usually already possible. For instance, if I want a septuplet covering two beats, I can have this notated with sixteenths by specifying a ratio of 7:8, or with eighths using 7:4. I can even notate it with seven quarters by specifying a ratio of 7:2, or thitryseconds using 7:16. However, this seems to have issues when used in conjunction with the "irregular full measure rest" trick. The specified ratio is not applied in a meaningful way, and in fact, I can create corruption pretty easily trying this. This seems a bug that presumably could be fixed without major architectural changes - but now we are definitely hitting upon the same territory as #202271: Copy-paste sextuplets/octuplets and their removal leads to corruption, which has proved a bit elusive.

In reply to by Marc Sabatella

Thank you for your reply. Very appreciated.

If one has a ratio of 3:5, with the 5-layer being 8th notes, the program - as far as I have seen - displays the 3-layer in 8th notes as well. What can one do to make it show quarter notes? Obviously same playback. Is there an option of 3:2.5?

In reply to by sami.amiris

I'm still not exactly sure what is going on when I try to specify 3:2 as the ratio, but anyhow, you can create the triplet of quarter notes in a 5/8 measure by specifying a ratio of 6:5 in the Notes / Tuplets / Other dialog. That is, create the 5/8 full measure rest, select it, Notes / Tuplets / Other, give a ratio of 6/5. This gives you 6 eighths or 3 quarters.

In reply to by Marc Sabatella

To me (I'm one of the core team member), MuseScore focus should be Conventional Western Music Notation. This is a fuzzy concept of course, but I believe the following points are not included in CWMN, a non exhaustive list, and so shouldn't be a priority at all.

* Irrational meters including 3/5, 13/666 and other oddities
* Tuplets above barline
* Polyrhythm in the sense of non aligned barlines
* And even, different time signature per staff, what we call local time signature in MuseScore 2

If these features become possible via a workaround, sure, that's fine by me. If someone manage to implement such a feature without making MuseScore more complex for a total beginner, fine by me. But if I have to choose between some code related to making MuseScore easier to use for 4/4 music, or review some code for one of the point above, my priority will always be the first one.

In reply to by sami.amiris

Advanced tab just solves one part of the problem, and we do use it in more than one place in MuseScore (advanced staff properties, customization of palettes, ...).
The other part of the problem is the technical debt accumulated by adding features that almost nobody uses. These features become hard to support and slowdown the development of more useful feature in the end.

In reply to by [DELETED] 5

FWIW, I agree "irrational" (non-quadratic?) meters are beyond the scope of what we need to directly support, and indeed it's not even clear what it means to support them. For non-aligned barlines and local time signatures, I think these *do* fall within things that are just conventional enough to want to support in some manner, but I'm fine with it being via manual workarounds. I'd say the same about tuplets over the barline. These are probably much more common than any of the other things, though, so if any of them were to have a native solution, that would be my pick.

In hindsight, our local time signature implementation is way too much code that comes with way too many restrictions and bugs to be worth it, and I suspect a simpler implementation that required more manual intervention might have worked out better, but oh well. If we do try to implement something for tuplets over the barline, I'd want to learn from the experience and maybe come up with a simpler solution that might require a little more work on the part of the user but that actually worked well and was not as much of a pain to maintain. not that I have any suggestions for what that might look like.

In reply to by mike320

I don't even know how to delete a triplet. I was confronted with this problem once again today and once again was faced with much experimentation. Finding out that what I entered was not what I wanted, I have no idea how to select it and then delete. My option was to just delete the entire measure and start over. I use Musescore but don't find it particularly user friendly and easy to use. I use it because it works on Linux. I honestly never recommend the program to anyone. Using it seems to be a form of masochism. I would say it is easier to use than Lilypond which is even more obscure. I find with both if you waste enough time, you can probably eventually get something acceptable, but you pay a heavy price. The bugs in Musecore are so numerous it seems fruitless to even try to list all the frustrations you encounter on a regular basis.

In reply to by Shoichi

That's really nice. How did you get the descending run which starts in the Treble Clef to cross over to the Bass Clef? Also when I try to copy your measure into my actual score, Musescore gives a message that "Tuplet cannot cross barlines." But this isn't crossing a barline although it is moving across staves. This second example (tertiary) is almost exactly what I would like, but I don't know how to replicate your work. The sample I'm submitting is your work, edited slightly.

Attachment Size
Tuplet Sample-quatrième.mscz 12.13 KB

In reply to by Shoichi

Well thank you. It was a great help, but it points up the need for much better cut, copy, and past operations in Musescore. One of the many bugs is that when you tie over from the previous measure, Musescore changes the first note of the run from tiny to full size. An unwanted behavior, many of which happen all the time.

In reply to by gBouchard

Feel free to start a new thread with suggestions for how you'd like to see cut/copy/paste improved. Right now it isn't obvious how they would have been relevant to this particular example. Also maybe start another thread explaining the difficulty you are having with ties. I couldn't reproduce any problem here - I deleted the tie into the run then added it back by clicking the note before the barline and pressing the tie button, and the note after the barline remained small.

In reply to by Marc Sabatella

Cut and paste is related because one the of persons, Shoichi, who reponded to my original querry or complaint created proper measure, but I wasn't able to copy and paste it into my score. It's always amazing to me that you can't find the weaknesses in Musescore which are so blatant. I already went over it fairly clearly in my original post. That's why I don't like to waste my time posting the bugs in the first place. One, there are so many, and two, the experts aren't able reproduce the wanted behavior that so many others experience. I just use it the best I can, but rarely is it a smooth experience where things can be quickly accomplished.
'

In reply to by gBouchard

No one is denying there is room for improvement. We make improvements all the time. But we can't fix problems no one tells us how to reproduce. You mention some sort of problem with copy and paste and another with ties, but I am unable to guess what are you are referring to as both work fine for me. So if you'd like us to actually fix the problems you find, you will need to start new threads, one per problem, and include sample scores and precise step by step instructions to reproduce each problem. Otherwise it may well be that the problems go unreported and thus unfixed. Posting precise bug reports is never a waste of time - it's the only way to get problems fixed. But complaining about bugs without posting precise bug reports most certainly is a waste of time.

In reply to by gBouchard

It's not totally clear to me how you did things that resulted in those extra rests. I am thinking you probably created two separate 16-tuplets, one for the top staff and one for the bottom staff, then entered notes into the first 12 slots for the top staff and the last 4 slots for the bottom staff, then hid the remaining rests, and also hid the bracket and number using the Inspector. Something like that?

If so, then basically, what you should have done is only create the tuplet in the top staff, fill it with all 16 notes, then use cross-staff notation to move those last four notes to the bottom staff (select them, Ctrl+Shift+Down). There then would have been no rests left to worry about hiding, and the result would have looked like what I assume you would have preferred to see (a single beam extending from top to bottom staff) and would also have required less effort.

It is actually perfectly possible to convert what you have into the preferred style I just described. To do so:

1) re-enter those last four notes in the top staff (replacing the invisible rests)
2) delete the bottom tuplet - probably easiest to simply select the whole measure and re-enter it, but for the record, to delete just the tuplet, you'd need to first select it, which is slightly trickier since you've disabled the bracket and number, but still quite easy - just click any note or rest within it, then click Tuplet in the Inspector. Now you can hit Delete to remove it. But you'd also need to clean up the use of voices, so ultimately you're better off just deleting the contents of the bottom staff for that one measure and just re-entering the quarter note.
3) select the last four notes of the top staff tuplet and move them down with Ctrl+Shift+Down, adjusting the beam as desired

As another note, probably you should have used the "Small" property of the chord, not of the individual notes. That way the beam and any other attached elements would also have affected. But maybe you wanted the full size beam. Anyhow, something to think about - normally you'd only use the "Small" property of the note rather than the chord is the intent is to show one small note within an otherwise full-sized chord.

Here's how things look if you do it this way:

tuplet-cross.png

In reply to by Marc Sabatella

That is what I did, because I don't know how to do things properly. That's why posted my poorly done sample. It couldn't be more clear that I often don't have any idea how to go about it in an efficient way. I'm mostly replicating somebody's existing work with small modifications. I'm just trying to get Musescore to look like the original notation.

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