extra ties and suspensions
Ever since I downloaded the latest version of Musescore 3 a few days ago (in Windows), the program keeps adding many nonsensical ties and suspensions to my compositions. I just opened one I worked on last night--it was fine then--and now it has as many as 20 such additions within 10 measures. I attached a pdf as an example; I didn't attach a Musescore file because, for all I know, it would be fine when you open it.
| Attachment | Size |
|---|---|
| error example for Musescore_KentLivezey.pdf | 38.46 KB |
Comments
In addition, it eliminated all the ties I had in the piece, so I just had to put them all back in. I'm reluctant to close the file and open it again, fearing all the mess will return. I'm also concerned about opening all of my other pieces.
Please also attach the score, without it there is no way for any of us to inspect what went south.
In reply to Please also attach the score… by jeetee
Thanks for the quick response. I didn't attach the score because I'm not ready to share it with the world yet. I fixed the ties and suspensions, so the score looks fine now. And, of course, my other compositions don't have this problem this morning. If it happens again, is there a way I can send it to you via email?
In reply to Thanks for the quick… by Kent Livezey
Okay, here's a score with some problems. When I opened the file, it was fine. A few edits later, these showed up; see bass clef in measure 15 and then lots of problems starting at measure 17. It seems to be random when these show up, so they may not appear for you.
In reply to Okay, here's a score with… by Kent Livezey
This sounds like memories of the unterminated slurs and ties that were being added back in version 2 for no known reason. I haven't seen anything like this in version 3. The ties have the marker indicating a certain duration but the end marker in the file is missing. @Kent, this isn't your doing so don't concern yourself with it.
The question is what is causing it. To you it seems random, but it's possible that you are doing something that causes it, but you don't see it until the next thing you do or you do something that causes a relayout of the score.
In reply to This sounds like memories of… by mike320
Hmm. I never tied together any of the notes in measures 19 to 23 as the score has, but I did tie the notes and chords between measures 20 and 21--but they disappeared.
In reply to Hmm. I never tied together… by Kent Livezey
As I said, this probably isn't your doing. It shouldn't ever create a tie without the end point being marked in the file, but that's not happening when the ties and slurs draw themselves.
So, what's causing the problem? What do these scores have in common? Were they created using the same template and is this a custom template? I don't know. Besides they are all on your computer, the problem ones must have something else in common.
In reply to Okay, here's a score with… by Kent Livezey
My only guess is that you selected two notes to tie rather than selecting only the first note, this is the only thing that would cause the results you see. Be sure when adding ties you only select only the first note (or add them in note input mode before even entering the second note).
But - was this file by any chance the result of import from MIDI? I could also see such problems showing up in that case depending on how the notes were overlapping in the original file.
In reply to My only guess is that you… by Marc Sabatella
I either input a note that spills from one measure to the next (thereby automatically making a tie) or I select only the first note and click that tie icon if I need to tie them manually (like when Musescore has been deleted my ties recently). I didn't know you could select two notes at the same time to tie.
I've never tried to tie notes across several measures. I don't even know how someone could play ties that connect notes several measures apart when many chords and notes are played with the same hand between the tied notes, so I sure wouldn't have tried to put them in there myself.
So there must be something else that caused these results.
I didn't import it from MIDI.
What's especially troubling is that a score can be fine, I close it, and then it's full of these ties the next time I open it. This has happened in several of my scores.
In reply to I either input a note that… by Kent Livezey
To be clear: you can't select two notes to tie them. If you select two notes, you get two ties, the second of which might look like these because MuseScore will have to hunt for something to tie that second note to. It's not something you normally do, certainly not for notes several measures apart, but the notation of tying non-adjacent notes is not uncommon for notating arpeggio-and-hold patterns, which is why the feature exists.
Anyhow, I can't say what might have happened here, but if I delete the ties, save, and reload, all is well. So if you can ever figure out steps to make it happen again, do let us know, but for now, that's what you would need to do to fix this one.
In reply to To be clear: you can't… by Marc Sabatella
After reading your message, I checked 9 of my pieces, and they were all fine (after I fixed the 2 that had problems earlier). I later opened one of them (not one of the ones with problems earlier), started making some small edits, and then about 30 of the strange long ties appeared. As before, most of the ties I had between adjacent measures turned into long ties that started at the same places but ended measures later. The score is attached.
So this has happened to 3 of my scores today.
In reply to After reading your message,… by Kent Livezey
Very strange. If you can catch it in the act - like seeing a specific command that takes a good score and turns it into a problematic one - then please attach the score in the good state and give us the instructions on how to mess it up. But I've never anything like this before. All I can think at this point is that the score already had some sort of strange invisible corruption that was just waiting for an excuse to make itself visible. But that still doesn't give me any particular insight into how multiple scores of yours are getting into that state to begin with. Is there anything you might consider unusual about your process of creating these?
When I run MuseScore under a debugger and load your score, I see some error messages printed to the console that do indicate a problem with the ties here. But it's hard to say where those problems can from. Are there any other you have that are similar but don't yet have the problem, so I can try one of those to see if there is anything suspicious lurking within?
In reply to Very strange. If you can… by Marc Sabatella
This is very clearly a bug because short of editing the .mscx file, the user cannot make a tie have only the start point indicated and not the end point. I am clearly not claiming the OP did this. Selecting 2 notes won't do this. Doing that will still give you two legal ties. The lack of the tie end in the .mscx file is probably the console error you are seeing.
As I told the OP, there must be something causing this, but I have no idea what it might be. The few times I've seen this in version 3 until now, it was actually version 2 scores but this doesn't seem to be the case here. We never did understand the cause of this in version 2 and until now I thought the problem was accidentally fixed in version 3. I do realize this is different because there are relatively few ties in version 3. In version 2 it was 10's of thousands being added. The similarity is that there is no termination in either case. I don't think this can be ignored.
If I were to look for the culprit, I would look at code starting in 3.3.2 that changed in 3.3.3 and start working my way backward from there. Since the version 2 bug was only occasionally reported I'm expecting similar results here, so the bug could have been introduced before 3.3.3, but I'd be surprised to find the bug has existed since 3.0.
In reply to This is very clearly a bug… by mike320
The OP wrote:
I either input a note that spills from one measure to the next (thereby automatically making a tie)...
Hmm...
Could adding a note of a (large) duration that "spills over" into a following measure with the resultant "automatic tie" produce a subtle "anchoring" difference from a tie entered the "normal" way? That is, by filling the one measure and then entering another actual note (i.e., not automatic) to be tied.
In reply to The OP wrote: I either input… by Jm6stringer
I do that all the time with no ill effects. In the current score I'm transcribing there is a long 3/4 section and there are a lot of 1/2 notes slurred to two tied 1/4 notes. I simply enter a half note, press s, enter the next note (which results in 2 tied 1/4 note across the barline). So, for example, I press 5 d s b s to get this and terminate the slur. I haven't seen any unwanted ties or slurs. I also routinely create notes I know will lead to a tie across barlines.
In reply to This is very clearly a bug… by mike320
Actually, selecting two notes and press the tie button will do this - try it and you'll see, MuseScore tries to start a tie at each note, so if the second note isn't followed by another of the same pitch, you get exactly this. But the OP said they weren't doing this, and indeed, the error messages I get on laod tell me these aren't "normal" extended ties, so I know they weren't created that way. What I can figure out is what could have causes this. Could be something involving time signature changes, that's abotu the only thing I can think of that changes how ties are attached. Or as I said, maybe the files were already corrupt. But indeed, it would be a bug, just not clear where.
The issues with multiplying slurs was totally different in the underlying code and the evidence that would be seen in the file, so even there is something that looks similar on the surface, it's not that.
It certainly could be something that changed since 3.3.2, but sicne right now only one user out of many thousands has seen this, and they have seen it on several scores, I think it's actually rather more likely the issue is inherent to something in common between those scores, meaning the bug has probably been around longer. Unless the bug is being triggered by some very unusual series of steps the OP is following regularly.
FWIW, though, there was a change to the behavior of the tie button for 3.3.3, https://github.com/musescore/MuseScore/commit/966ddc23a09669c48ededa902…. This could potentially explain problems adding new ties, but it wouldn't affect existing ties. I see this commit that deals with copy/paste of spanners: https://github.com/musescore/MuseScore/commit/49fedc47c409dc57039ca5b53…. This commit deals with layout of slurs but might affect ties: https://github.com/musescore/MuseScore/commit/e5d93481e0d6f28efcfca9b2e…. This one is about ties on unisons (same chord): https://github.com/musescore/MuseScore/commit/6638eaf980664f6967dc68d16….
I guess we'll see if anyone else reports anything like this...
In reply to Actually, selecting two… by Marc Sabatella
Selecting two notes, will lead to 2 ties, both of which are normal, though the second one will be unwanted. Their terminations are still entered into the .mscx file even though it may be 3 measures and 24 notes later.
Otherwise you have restated what I have already said and listed changes between 3.3.2 and 3.3.3. As I indicated, I'm not convinced any of these will be the culprit. In an earlier post I indicated that the action that leads to the problem might not show up until some later time. Until the bug gets reported more, all we can do is ask the OP what he's doing on these scores. It sounds like it's happened to him more than once, so with his cooperation we might be able to figure this out.
In reply to Selecting two notes, will… by mike320
If you check the score, you'll see that all the ties are exactly of that form - they are exactly the sort of ties that would be created if the start and end notes had been selected n just the right way. Each tie connects two notes of the same pitch, just not two adjacent notes, So it was within the realm of possibility. But, this wouldn't lead to the error messages I saw on the console, so I do think we can rule out this having somehow happened accidentally.
In reply to If you check the score, you… by Marc Sabatella
I now see that the termination point of the tie is being reported, so not the same symptom as the version 2 scores.
What's strange is that the tie get broken when it end is on a different system than the original. It's drawn to the end of the system, but looks like it's drawn from the previous note/rest at the destination note. MuseScore recognizes the notes are tied and moves them as though they are tied if you repitch one of them.
In reply to I now see that the… by mike320
I appreciate the attention you folks are giving this. Hopefully this can be fixed. I didn't have this problem for the first 1-1/2 years I used your various updated programs; it all started with the most-recent one, so there must be something in the new update.
Marc asked if I could send some other scores that haven't shown the problem, and see if you can catch it in the act. I'd be glad to send some. May I do so via email? If so, please send me a message (kentbl@gmail.com) and I'll respond with a half-dozen scores. After opening, I suggest you make some small edits to the first dozen measures and see if the ties appear. That's what happened to me in the 3 scores yesterday.
In reply to I appreciate the attention… by Kent Livezey
Better to post a couple here, there is not likely to be anyone person who will be the one to devote time to this, better to have multiple people each try it a but.
In reply to Better to post a couple here… by Marc Sabatella
Rather than send new scores that might have problems, attached are scores from the two I sent earlier that did have problems, but from the versions just before the ones I sent last time. Once you open them in the new Musescore and save them, I'm thinking they should have the same the problems as their most- recent versions. But you may first be able to see what these scores have in them that triggers the problem. It's worth a try, at least.
In reply to Rather than send new scores… by Kent Livezey
What kind of things where you doing when you noticed the extra ties?
In reply to What kind of things where… by mike320
Just very small edits, like adding a fingering or changing a note. Then, probably, saving them. I can't remember if I saved them before all of the ties showed up.
In reply to Just very small edits, like… by Kent Livezey
I just changed a time signature in another score, and lots of ties appeared. So I closed it without saving it and tried the same change, but the ties didn't appear. So strange.
In reply to Rather than send new scores… by Kent Livezey
That's great you have these, files we know are potentially susceptible. So far I haven't managed to trigger the problem in a few minutes of experimentation, but I'll keep fiddling. Meanwhile, let us know if you ever figure out which specific edit seems to be triggering it for you.
In reply to That's great you have these,… by Marc Sabatella
You might try changing time signatures.
This one just worked 3 times in a row. Open the attached (which has no long ties) and change measure 8 to 3/4 time. Many long ties should appear below.
In reply to You might try changing time… by Kent Livezey
Confirmed
In reply to Confirmed by Jojo-Schmitz
Same thing happens if you change measure 17 to 3/4 time.
In reply to You might try changing time… by Kent Livezey
Thanks for that - this one I can reproduce. Ties after a time signature change can get messed up indeed. See #277800: Adding time signature breaks a tie across barline, which is related but this is a new wrinkle that is rather worse.
It seems any time there is a time signature change that involves adding new beats (because the region being affected is not evenly divisible according to the new time signature), then ties after the changed region get affected in this way.
I will submit a new issue on this.
I suspect that your other cases also involved time signature changes, since that's always the one type of edit that can involve the need to rewrite ties.
One workaround for now is to not change time signature in a way that requires extra beats like this - insert extra measures as needed so it comes out even.
In reply to Thanks for that - this one I… by Marc Sabatella
Okay, I'll try that workaround for now, and see if it works. Please let me know when you've addressed this. Thanks to all of you for the prompt help.
In reply to Thanks for that - this one I… by Marc Sabatella
FWIW: I'm still using:
OS: Windows 10 (10.0), Arch.: x86_64, MuseScore version (64-bit): 3.2.3.7635, revision: d2d863f
...and can not reproduce. Here's what I see after changing bar 8 to 3/4:
So maybe something changed leading to 3.3.3?
In reply to FWIW: I'm still using: OS:… by Jm6stringer
or 3.3.2, which is that the OP seems to be using?
In reply to or 3.3.2, which is that the… by Jojo-Schmitz
If I'm the "OP", I'm using whatever version came out about 5 days ago, the day this problem started.
In reply to If I'm the "OP", I'm using… by Kent Livezey
Yes, you are, but the score stems from 3.3.2 I thought to have seen and got created some time in August?
In reply to Thanks for that - this one I… by Marc Sabatella
I have submitted the issue as #297957: Ties extended in region after time signature change.
The history of the file turns out not to be relevant. My analysis gives me a pretty good understanding of the cause and a couple of ideas about solutions, but I'm hoping someone else takes it from there,
In reply to You might try changing time… by Kent Livezey
From what I've seen here, when the time signature is changed, the tie jumps to the next same note, not the one to which it belongs.
In reply to From what I've seen here,… by Ziya Mete Demircan
I started writing that score on August 12. The first problem arose last week, right after I updated the software.
In reply to I started writing that score… by Kent Livezey
Ok, that history may play a role, or may not...