Request alternate position key signature when using begin repeat sign

• Mar 12, 2019 - 15:12
Reported version
P1 - High
S5 - Suggestion
  1. key signature change during score.
  2. add start repeat sign
  3. key signature shows before the repeat sign in the prior measure. Should be after bar line change.

This has happened in multiple scores and parts. I am sending one for your reference.

I am using Version

Attachment Size
1st Baritone (BC).mscz 38.81 KB


According to Elaine Gould in Behind Bars p. 234, the repeat marks should be placed after the key signature. MuseScore 3 does this correctly.

I agree with the OP! And it's impossible to fix because removing the check from automatic placement does not allow you to move the key signature to the other side of the bar line. Gould at times has an agenda to change notation rather than explain what has happened in the past. This is a obviously a case.

In reply to by mattmcclinch

Sorry but I respectfully disagree! All printed music I have worked with always has the key signature after the barline. Whatever type that may be. I also found another problem when using the repeat bar lines. At the end of the repeat section, all the empty measures in the repeat do not multi measure. I have enclosed an updated part to show this problem. It is at rehearsal marking "F".

Enclosed original part (PDF) which I am transposing from Treble Clef to Bass Clef for a Baritone player in the band to show how it is printed.

Attachment Size
1st Baritone (BC).mscz 42.89 KB
1st Baritone Bb.pdf 1.17 MB
Title wrong position for keysign when using beginnng repeat sign Request alternate position key signature when using beginnng repeat sign
Severity S3 - Major S5 - Suggestion
Priority P1 - High

I've seen it both ways. I think the layout Gould recommends is more logical in most cases, after all, the key doesn't normally change again on the repeat. The only time that would make sense if there is in fact a second key change within the repeat, so the key signature is needed - both at the beginning and a courtesy at the end - within the repeat. Which to me says, either should be possible. And in fact, it already is possible but only with some fairly convoluted workarounds (eg, marking the real repeat invisible, adding a fake one in the previous measure). So, I do support implementing a simpler option, such as perhaps a property to set on the key signature.

You may not need to change key or time signatures, but it also doesn't hurt if you declare the same ones again. So, technically, only repeat-bar first is needed. But I'm all for more options.

I do have one thing to add, though: the behavior I've seen most often is the one in MuseScore 2. There, the repeat bar usually comes first, except at the beginning of a line. Then it comes after any clef, key signature or time signature. I can't recall ever seeing the repeat bar before the main clef of a line.

I would prefer not to have to manually set a different repeat symbol depending on where it falls on the page, as that can change during composition. The behavior in MuseScore 2 was perfect for my needs.

To be clear, we're not talking about repeats at the beginning of line, but ones in the middle, and how they are positioned relative to a change of key that appears within the line.

In reply to by Marc Sabatella

Yes. Like files I sent with my request. No at beginning of line. Key chnge is at therepeat sign. It is to right of single barline, but moves to before for any other barline. This happens in the file I sent at rehearsal "E". Also after last update, time signatures are showing twice. When line break is used or chahged. I have been setting one to invisible as a work around.

If you are having some sort of issue with time signatures, please open a new issue and attach your score and most importantly, steps to reproduce the problem.

The issue with the multimeasure rest is also unrelated, but this one is known and already has a fix slated for a coming update - see #282234: Volta breaks multimeasure rest

Again, for key signatures, what we currently do is correct "by the book", but indeed, an easier way to reproduce the look of historical scores that the standards of their day would be nice.

Any updates on this? I notice it missed the 3.1 inclusion, despite the high priority. Is there some sort of holdup where input from users might be useful, like deciding how best to implement this?

I think the easiest implementation would probably be to add an additional repeat to the palette. This would have the least effect on preexisting scores that are already okay with the $50 book form. It would also bypass a need to create measure-level and score level (Style->General) level changes, and avoid the issue of which version should be the default. People can just use the one they like better.

The main issue would be differentiating their icons. I would suggest putting a sharp sign before the start repeat for the current version, and a sharp sign after the repeat for the new version. Make the sharp sign gray (or checkerboard pattern) to show it's not a real key signature. (I chose a sharp sign because a flat sign would be close to the dots of the repeat sign.)

The other issue I can think of is importing scores from version 2 or lower. I would suggest keeping it where it looks the same, rather than forcing a change on them. My reasoning is threefold: (1) it's generally best minimize changes when upgrading versions and (2) key signature before the repeat still seems the more uncommon choice (3) no one complained about the previous behavior, so people are less likely to complain.

Finally, would it be too much to ask to actually quote the section of the book that establishes that key signatures go before the start repeat bars? I wanted to double check that it was being interpreted correctly, but $50 is a steep price to pay to be able to verify an implementation. I still find that the way it is currently implemented just looks odd, and I wonder if something got missed, like maybe there is supposed to be a barline before the clef.

Because no barline before a clef usually means that it is a courtesy clef, not a full clef.

New features are always lower priority than bug fixes, even for the same setting of the "priority" field. I have no doubt this will get looked at in due course, though.

Here is the passage from Gould:


Title Request alternate position key signature when using beginnng repeat sign Request alternate position key signature when using beginning repeat sign

Fix typo.

In reply to by Marc Sabatella

Then it is as I suspected. It says the start-repeat sign should come after a clef, key, or time signature change. It says nothing about not having a regular bar line before the clef, key, or time signature.

I would suggest this is what looks so wrong to most of us. Key signatures and time signatures always come at the start of a measure, not at the end of the previous one. It looks inconsistent for them to appear before the bar line.

This could be fixed by adding a regular bar line before the key and/or time signature, while also keeping the start-repeat sign after.

I also note that this bug isn't a normal feature request. It is a regression. The desired behavior existed in MuseScore 2, but does not exist in MuseScore 3. It is just that the proposed fix involves adding a new feature.
A feature request does not normally block user from upgrading.

It was a bug (according to Elaine Gould) in MuseScore 2 which got corrected in MuseScore 3. Making it an option in MuseScore 3 is a feature request

Sounds then like you are actually requesting a feature different from what the one actually being discussed here, then, The original post states "key signature shows before the repeat sign in the prior measure. Should be after bar line change". It's true the key signature shows before the repeat sign, but as Gould discusses, it is just plain wrong to say it "should be" after. If one wants that non-standard placement, it's a feature request to add it.

Sound like you are talking about something else entirely. That is, you agree the key signature (or clef, or whatever) belongs before the start repeat. And in the case where the preceding bar has an end repeat, you are also agreeing with that (unlike some others in the discussion who object to it being placed between the end and start repeat). You are now requesting a separate new feature, where the clef/key/time signature still goes before the start repeat, but unless in traditional music notation, a new barline also gets inserted in front of it. I can't say I've ever seen that as far as I can recall, but feel free to post examples/quote in favor of that additional new feature. Probably best to start a new Suggestion for that, though, and keep this one focused on the original topic of being able to move the clef/key/time signature after the start repeat. The separate new feature request is, "Insert additional barline in front of clef/key/time signature when occurring before a start repeat".

It is not true that MuseScore 2 inserted such a barline. It simply placed the clef/key/time signature after the start repeat barline, which is, again, wrong, which is why that bug was fixed.

So again, there are two separate things being requested here:

1) an option to move the clef/key/time signature after the barline, like MuseScore 2 erroneously did by default
2) an option insert a new barline in front of the clef/key/time signature while still keeping it before the start repeat

Title Request alternate position key signature when using beginning repeat sign Request alternate position key signature when using begin repeat sign
Frequency Once Few

I've now encountered the time signature splitting a double repeat. I'm transcribing early music and it seems that in 1571 they didn't follow Elaine Gould. I'd like to see an option to locate the time signature to the right of a repeat barline. See snippet showing original and MuseScore versions.

Attachment Size
splitdoublerepeat.jpg 8 KB

In reply to by John Kilpatrick

I'll second that request for an option to place the key/time signature to the right of the repeat sign. I, too, am transcribing early music (though not as early as previous commenter). I find the split repeat sign to be very disconcerting. In over 50 years of playing I've never encountered it in any published score I've seen.

Workaround No Yes

I have a work-around. Insert a measure before the key change and place the repeat sign at the beginning of the inserted measure. Change the inserted measure's time signature to 1/64 and place a 164 rest in it. Now, make the time signature, inserted rest and end bar line invisible. Finally, select the repeat sign and use its position X offset to adjust its location relative to the key change depending on how many accidentals there are. It should be noted that this will alter the playback by a 1/64 rest.

Attachment Size
KeychangeFix.png 17.19 KB