Ties not correctly formed when adding an interval

• Aug 28, 2014 - 09:44
S4 - Minor

The two bars were created firstly with the top notes tied together. Chord tones were then added to the first note which produces ties, although not yet having anything to tie to. Harmony notes were added to the second bar (same voice), but the ties do not extend to join them. The second chord also sounds, so is clearly not seen as tied.

This process seems to work correctly with notes other than whole notes.

I haven't attached an mscz because it seems repeatable behaviour.

(But V2 is a great improvement on V1 so far!)

Attachment Size
Broken-ties.png 27.68 KB


I've just found that the correct behaviour results if you first create the whole chord in the first bar then use the tie (+) to create a note in the second bar and add the subsequent notes (from the keyboard, at least). The ties are not formed at this stage, but if you then select the whole of the first bar and press + all the ties form correctly.

It should work either way though.

Why have you closed it? The behaviour is still wrong. Why should I have to decide on my harmonies and chords as I go? It should work whichever way you wish to work.

Status (old) closed needs info

I'm confused as to what you mean here. First you create a chord of two whole notes, then tie it to another chord of two whole notes. Then you add another note to the first chord - how? You say you see an incomplete tie on that note, but you shouldn't see *any* tie unless I am misunderstanding something. Adding a note to a tied chord does *not* add a tie, although that would make a good feature request. Yet you picture does appear to show some sort of incomplete tie. I just have no idea how you got there. Could explain what you did step by step from the beginning?

Hello, Marc, obviously my explanation isn't as clear as I thought. I'll try again.

1. Create a whole note in the first bar. (Although I've now found this is the same for any length of note).
2. Extend the note into the second bar by pressing +.
3. Select the first note and press Shift 6, or whatever, to add a harmony note. This note will have a tie hanging from it. Adding more harmony notes will have the same.
4. If you try to extend the harmony note into the second bar by pressing +, then you will delete the original tied note in the second bar. So you need to select that note instead and add the harmony notes.

At this point you have the situation in the first diagram and no way of connecting the vestigial ties.

I hope that is clear now.

Status (old) needs info active

Ah, Shift+6! It works as expected (expected by me, anyhow) if you add the note using Shift+letter rather than Shift+number. And by "as expected" I mean, no tie is created at all when you press Shift+letter. MuseScore has never created ties automatically when adding notes to chords, although as I said, that would make a good feature request. So the bug here is that you get the disconnected tie in the first place.

If you add a note or interval to an existing and tied (one-note-)chord, it should add tied notes to that chord.
Whether 1.x did it or not ;-)
If you don't want that, use a different voice

As I said, it makes a nice feature request. But it is complicated. What if only some of the notes in the chord were tied? What if the chord that follows already has a note there? What if that target chord itself is tied (perhaps only partially) to the next chord? Automatically creating ties is a problematic subject. So, not that it's not worth trying to handle some of the easy cases, but again - it's a feature request.

Meanwhile, the bug where a disconnected tie gets creates should be fixed; that isn't doing anyone any good.

Interesting how different people work, isn't it? I add the interval I want, not the note name I want. I hadn't even thought if doing it that way. Never occurred to me it might work.

But the behaviour I'm seeing is different to 1.3, so I assumed it was trying to work 'correctly'.

As to whether you would want some notes tied and some not, it wouldn't make sense to have them in the same voice. That is illogical, and when I've wanted that then I use separate voices.

So whether it is a feature request or a bug that needs fixing and making consistent when adding intervals or note names is for yourselves to consider. My preference would be for it to be fixed and made to work.

MuseScore has long supported chords where only some notes are tied, we can't take that feature away now, even if I agree that most of the time someone does that, they probably should have used a separate voice. And yes, the behavior is different from 1.3 - and that difference is a bug I think should be fixed. As I said, the disconnected ties don't do anyone any good.

As for the new suggested feature where adding notes to chord that contains one (or more, or all) tied notes would automatically also create new notes in the tied-to-chord (assuming all the notes were tied to notes in the same chord!) and then automatically tie them as well, at least where possible - I do think it would be nice. It doesn't seem quite as important to me as the abiltiy to create tied chords directly in the first place while in note input mode - that's bugged me for years! - but maybe both features could be added together at some point.

I did implement a feature of trying to automatically add tied for the new "repitch" mode, so I do have some idea what would be involved in making that happen here too. it's different, because in repitch mode, I could just follow the existing ties, but it was still not something that could be made to work "as expected" by everyone in all those various cases.

Anyhow, I like the feature request, but right now of course we're more concerned with fixing bugs to get 2.0 released than in implementing new features, so it will be kind of low priority.

Well, a tied note has a total length of the addition of the tied notes' length.
So should an added note to that chord. If something is in whe way, it should get overwritten, just like when you overwtrite a shorter note with a longer one.
No chord should ever be only partly tied anywhere, if you need that, use voices, so I see partly tied chords as a bug.

People use that all the time, though, so we pretty much have to still support it. You can even find it in published literature. We also support ties to non-adjacent chords, mostly to be used for things like arpeggio notation (see http://musescore.org/en/developers-handbook/scrapbook#Extended-Ties). And we at least partially support one-side ties, for things like ties into or out of voltas and repeats.

Like I said, I actually implemented something sort of like this for repitch mode, so I do have some idea of the complexities involved. And I'm not saying it isn't a good feature request - I'd like to see it to. I'm just pointing out that it is not quite the no-brainer you might at first think.

I'm having trouble getting notes of a chord to tie across a bar line.

In this case, a chord of two notes, a crochet to a quaver in the next bar.

I put in the first two crochets, then pressed shift + and got a tie to the top note only.

I put in the notes as dotted crochet - and it put in the second note only, with a tie at the top.

If I select the first note and put in a slur, it puts it at the top of the notes, above the tie.

I have tried selecting both chord and pressing shift + and it still only wants to tie the top note.

Your advise appreciated!


Thanks, Jojo, very helpful.

Now I can mostly put in the ties I want; however, another problem has arisen.

On attempting to tie the bottom note of two chords, an F#, the tie wants to jump the whole bar and connect to a different F# in the next bar.

There are several of these tied notes in the piece, most of them work, the problem only seems to arise on the bars where there is an F# (it's in the key of C).

I'm using Musescore 2.0 on Windows 7.


I would suggest you start a thread on the Technology Preview forum, posting the score you are having trouble with and precise steps to reproduce the problem. I'm guessing it will turn out to be something you aren't doing correctly, but it's hard to tell for sure. In any case, it's not connected to this issue, so let's discuss elsewhere, and if turns out there is another bug, we can file that separately.