Appended measures in different time signature without being marked
I'm working on a 3 movement sonata. The second movement was in 2/2 (or cut common) and now I'm half way through the third movement, which is in 3/4, and after I have appended measures, the time signature has reverted to 2/2 without it being marked in the score. Any ideas how to fix this?
https://musescore.org/en/node/81031 It's the same problem as this one, except that I didn't play around with the nominal/actual durations.
Thanks in advance.
Comments
If it helps, I've attached the score below. It changes to 2/2 in 245. I've changed the "actual measure duration" just for that one, but I don't want to have to do that manually for each bar from now to the end. I can actually remember this happening before, but I don't remember how I fixed it...
Edit: actually, I closed the file, reopened it and the problem was gone. I did, however get a notice saying that my file was corrupt, but it opened anyway.
In reply to If it helps, I've attached by Kaspar
Sorry, I spoke to fast. When I append more measures, they are back in 2/2!. I managed to get to the end by manually changing the actual measure duration for each bar, but I'd still like to know how to fix this problem in case it happens again.
In reply to Sorry, I spoke to fast. When by Kaspar
I suggest you look at the pop-up warning that appears when you try to open a corrupted file, and determine which measures are corrupted. The most dependable way to fix metric corruption of this sort is to delete the affected measures entirely (highlight the measure in a blue box and hit CTL+del), then insert new blank measures to replace them, and re-write those measures manually.
Then save the file under a new name, and work from that to complete your composition.
As to how this happened, I don't know, but I suppose it's possible the program might read the INITIAL time sig (that of the first movement) and apply that automatically to any appended measures. The actual corruption may have occurred as a result of attempting to 'fix' the problem measure by measure in the measure properties dialogue. I'm only guessing, of course, but instead of doing that, the way I would deal with this sort of thing is, BEFORE entering any new music to those measures, try to delete the time signature at the start of the appended section. If that doesn't work, apply the time signature you want from the F9 palette to the same measure, overwriting the auto-generated one, and then make it invisible. And if THAT doesn't work, delete the entire measure with the unwanted time signature in it (same as above: CTL+del).
As the message says, the score* is corrupt, so if you have an insight into how that might have happened, that would be useful - we do try to fix any reproducible cases of corruption we can.
In your case, measures 89 to the end (measure 94) have full measures in them that are inexpliciably set to only 3/4 instead of 4/4. Since they are empty, the easiest workaround is to simply delete the measures completely (Ctrl+Delete) and insert new ones.
*The score attached in the thread you referenced. I guess that isn't your score, but I don't see that you attached one. If you'd like further assistance, please post the score and precise steps to reproduce the problem.
In reply to As the message says, the by Marc Sabatella
Sure. Here is my score. I took it down when I thought I'd fixed the problem.
I looked at the list of corruptions when I got the notification upon opening the file. They all seem to refer to the appended bars being in a different key signature. (eg. "Measure 246 Staff 1 incomplete. Expected: 3/4; Found: 2/2") I should also add that any bars I append or insert into the last movement are in 2/2. Strangely, the 1st movement is in 3/4 - 2/2 is the signature of the second. I also tried deleting the problem bars and this removes the corrution, but then when I try to add more, they appear again in the wrong key signature. As I said, I'm at the end now (I just need to add the other parts, which shouldn't be a problem), but it's a strange problem isn't it?!
Thanks again.
In reply to Sure. Here is my score. I by Kaspar
Corruptions are always strange. In this case, the problem is with the apparently empty measures; they have full measures rests that are somehow set to the wrong number of beats. Not clear how that happened. Probably best to copy the good content to the clipboard, delete the problem measures, add new ones, then paste back the content.
In reply to Corruptions are always by Marc Sabatella
Thanks,
The problem I was having is that any measures I added would be in 2/2, even after deleting the problem measures. I've somehow worked around it by deleting the 3/4 signature and then putting it back again. Then all measures after the 3/4 time signature are in 3/4, as they should be.... unless I then add more measures after this process, then they appear in 2/2.
Thanks again.
In reply to Thanks, The problem I was by Kaspar
If you can find precise steps to reproduce a problem in a score not already corrupt, please let us know, so we can fix the bug.
In reply to If you can find precise steps by Marc Sabatella
Well, I think I have reproduced the problem by simply recreating the basic layout of the piece I am working on, with the same time signatures and actual/nominal bar lengths where necessary. Here is how:
1. begin with 3/4 time signaure
2. after a few measures, add final barline, then 2 cut common time singature to begin movement 2.
3. in the first bar of cut common, make the first measure 1/4 (measure properties - actual: 1/4, nominal: 2/2) as an upbeat.
4. after a few measures, add an end-start repeat with a 3/4 measure to the left and a 1/4 measure to the right (3/4 to make up for upbeat at beginning of movement, 1/4 for upbeat to second half).
5. after a few measures add end repeat with a 3/4 measure (nominal 2/2) as final measure of movement.
6. add 3/4 time signature.
7. ctrl+b to add new measure - new measure is in 2/2!!
Is that enough information? I've attached my experimental below just in case I haven't explained myself well enough above, you can see how I did it.
All the best.
In reply to Well, I think I have by Kaspar
One other thing.... I deleted the nominal 2/2, actual 3/4 measure (measure 21, created in step 5) and then added some new measures and they are correctly in 3/4. So that is the problem, but is it a bug or something I have done wrong?
PS after step 7, I closed the document and got the corruption notice upon reopening.
In reply to One other thing.... I deleted by Kaspar
As far as I can see, when you added the 3/4 time signature you were adding it to a measure that was nominally 2/2 but which you have "changed" to an actual duration of 3/4. When you then add extra measures they take on the nominal time signature of the final measure, not its actual new duration.
This only appears to happen when you manually alter the actual bar duration of the measure and then add the time signature that matches that duration. Although I haven't extensively tested, it appears that if you set a bar duration and then add a different time signature then subsequent new measures will take on that time signature.
Possibly to do with coding and with the different variables used to store measure duration?
In reply to As far as I can see, when you by underquark
That also seems correct - I made the previous bar both 2/2 nominally and actually, and then appended a few measures. They were correctly in 3/4. Why then does this happen? It's not how it should work is it?
I would think that when I change the actual duration of a measure, it would only affect that one measure. If I add a 3/4 time signature to the following measure, shouldn't that time signature cancel the previous one and remain until the end of the piece, or at least until the next change of time signature? That's what happened, actually, it's just the appended measures which revert to the previous time signature. If I append a measure, I would expect it to be in the duration of the current time signature, not the previous one.
In reply to That also seems correct - I by Kaspar
If something results in corruption, it's a bug. Could you report this officially to the issue tracker (eg, via Help / Report a Bug, from within MuseScore).
In reply to Well, I think I have by Kaspar
Confirmed MuseScore 2.0.3, rev-3c7a69d
running Windows 10