Figured Bass Continuation Lines Disappearing

• Sep 2, 2019 - 20:28
Reported version
S4 - Minor

1) Create a score utilising figured bass.
2) Make sure that there are some continuation lines in the figured bass to experience the issue.
2a) Issue seems to happen with both extended and non-extended continuation lines.
3) Save the score. Exit the Musescore program.
4) Reopen the Musescore program. Load the score.
5) The extended continuation lines directly following a number figure are no longer visible.
6) The isolated continuation lines are completely missing.

In general, for someone who works with a lot of figured bass in my scores, I've often found continuation lines to be somewhat clunky or messy to implement. I feel like I have to spend a lot of time on just those continuation lines alone in respect to filling out an entire figured bass.

Let me know if I can assist in any way to explain the issues I've been experiencing. Primarily, I'm mostly concerned about the issue of disappearing figured bass continuation lines. (The previous paragraph is just a general note in respect to the same feature). I can't seem to find any way to work around this.


Then you post it online, and there is a horizontal strike from the left margin to the place where the continuation lines are discontinuo. I have suffered this for a while, I guess I neglected to report it. The workaround is to use staff text to place a __. "That's why they call it a work-around".

In reply to by mike320

In that issue, they're reporting that the duration line is hard to see, but still technically visible.

My issue is not to do with the duration line that you see above the number figures, but the continuation lines that would appear in place of the numbers themselves.

Just in case there's some confusion here in my last response, I was referring to mike320 marking this issue being a duplicate of #282893.

I believe this issue and mike320's referral as (though somewhat similar in some minor respects) being two very separate issues due to the mechanics, features and visuals involved. So, I marked this issue back to "active" rather than "duplicate".

I agree figured bass is rather glitchy. Many of the issues are regressions from MuseScore 2, the result of that part of the code not keeping up with the changes in the core, but some may have been around since the beginning.

Has anyone tried the open source Figurato font ( From what I have seen it's pretty amazing. One could almost get away with using this, perhaps in conjunction with the new Roman numeral facility to get the navigation. Except, I guess, for the continuation lines, that's maybe the one thing we still would probably need to do ourselves to do right. But maybe this is a better way for us to go long term given that none of the active developers really know much about figured bass or about this area of the code?

In reply to by Marc Sabatella

Interesting! Is the "new Roman numeral facility" an upcoming feature? I've wanted to explore Roman Numeral Analysis in MuseScore previously, though realising that there was no technical feature for this, I haven't put any further thought into it. Something like manual entry of Roman Numerals while specifying durations could be quite beneficial in general, let alone improving the functionality of figured bass, as I imagine the coding of duration and such to be rather similar. The only difference would be representation in figures, obviously (especially the implementation of continuation lines). I suspect that Roman Numeral Analysis would be more widely used by users of MuseScore too, since figured bass is more of an early music enthusiast's interest.

Either way, any feature that can improve upon the current implementation of harmonic/chordal representation and analysis for all styles of music would be much appreciated, I think.

Also, I wonder, do the continuation lines in lyrics share the same code as that in figured bass? I feel I've also had similar frustrations in the implementation of lyrics. However, the continuation lines here seem to remain visible in saved scores, unlike the current issue with figured bass.

In reply to by Erisceres

The Roman Numeral Analysis (RNA) feature is in 3.3, whose Beta is now available. All it really is is a new line, like lyrics, chords, figured bass, etc., and a special font that allows you to say "bVII642" etc and get the usual Roman Numeral representation/figure stacking almost all the time (i.e., it does not perform musical analysis, etc.). But that's plenty, IMO.

In reply to by Erisceres

Regression No Yes

This issue escaped me completely, sorry (but I have been on "leave" from MS for a while...).

I had a look at the source for the FB: the problem with continuation lines is that someone changed the way they are stored in the files, but did not converted them back when reading or did not update accordingly the parsing algorithm. So, the two parts (writing and parsing) do not agree with each other.

It is probably a rather simple thing to fix, as well the other issue with duration lines.

I would like to know which problem you are having with entering cont. lines; I would think it is a rather straightforward process, but I can be biased.

@Marc : surely anything could be improved, but I do not understand how a different font could provide a significantly better result, as:

  1. none of the current issues with F.B. is a problem of glyphs or font
  2. the current font already contains all the needed glyphs (and a few more, IIRC)
  3. F.B. layout is quite more than shaping the correct glyphs: it is mostly a matter of non trivial horizontal (and occasionally vertical) alignment, which a font cannot provide by itself.

EDIT: after some debugging, I think the problem is with fractions: the duration of the F.B. item is written out as a fraction, but it is read back as an int, so the actual duration of the item is bogus and all lines are random (actually mostly null, as the duration becomes 0 in most cases).

@Miwarre not being a FB expert, I don't have a good sense for what issues might still remain to be discovered. But my concern was that given that none of the rest of us understand the system (FB itself, or the code involved) it might prove difficult for us to maintain going forward. So it was just another option to consider. I'd rather, though, there be someone who can maintain the code we already have, so it's good tosee you back!

But FWIW, have you tried Figurato? I t actually does quite a bit of alignment itself, just as my Campania font for RNA does (but Figurato is far more sophisticated). I encourage you to check out the link - even if there is no reason for MuseScore to use this, I think you'll find it very interesting.

Regarding the issue at hand - you probably (?) know that ticks were replaced with fractions by Werner at 3.1. For as huge a change as it was, it went surprisingly smoothly, I think we can count on the fingers of one hand the number of regressions that were found as a result. But this is exactly of the sort of issue I'd have expected to see more of, and won't be at all surprised to find this is the cause. The other possible candidate, though, is a change Dmitri made a couple of years ago to how the file format was written, to make it less dependent on numeric id's for things that links and connecting spanner start/end, also to remove explicit indication of most measure numbers - the ultimate goal being the format is now much more "diff" friendly (and ultimately resulted in View / Score Comparison Tool). I could also easily believe those changes somehow missed the continuation lines.

Hi Miwarre. Thank you for looking into this. In response to your first post, the issue I've been experiencing is that any entered continuation lines have been missing upon reloading the score.

The other issue, the clarity in duration lines, is an aesthetic issue that has not personally affected me, but I understand that they can be difficult to see.

While we're on this topic of fixing and improving FB, I might raise one point, again to do with continuation lines (or more so vertical alignment). As a continuation line represents a sustained note between separate figures, sometimes only one of the figures (as in the number/glyph) is meant to be sustained until it feeds into the next figure. Let's say i have a figure under a note as just "3". Under the next note I might want "6" and "3", but I've decided that the "3" from the first note continues into the second, making use of the continuation line. However, the "3" under the first note aligns with the "6" of the second. It would be nice if I can enter vertical white-space (aka a newline) inside the figure to force the "3" to nicely align under the "6". I can't seem to be able to do that at the moment. Instead, I have to manually adjust the whole figure's alignment until it seems to visually line up. I think it would be nicer, quicker and more practical to just put a newline in the figure as most of the alignment is already automatic. Does this make sense?

For some reason, the PR was not merged yet and it will not be in ver. 3.3.1 due tomorrow or so. I hope it will be in the next release.

I think the original 3.3.1 list is more like 3.3.2 now, 3.3.1 is more of an emergency release just for the people with missing palettes or crashes on startup. And you'll be pleased, perhaps, to know that part of that fix is to revert to Qt 5.9 on Windows.

But since the % issue is a pretty critical regression - really the only other one I can think of from 3.2.3 to 3.3? - I do hope a fix is found for that and included in 3.3.1 as well.

In reply to by Marc Sabatella

Telling me gets only a tearful shrug of sympathy. Is Anatoly on board? I have mixed feelings about the 5.9 deal -- I would rather have heard "5.12 for Mac" -- but what about the people who need the Accessibility fixes in 5.12 (win)? As Jojo has most effectively argued, they haven't any choice (while the OS X 5.9 users would merely have to second-mortgage their houses to buy newer Macs were that to change to pass).

Apparently 5.9 is fine for accessibility, the issue was with 5.12.3, which Windows has been using for quite some time. Moving forward to 5.12.5 fixed accessibility, but so does moving backwards to 5.9. Apparently the issue was only with the releases in between.

Status PR created fixed

Fixed in branch master, commit c81fa7300a

_Fix #293998 : Fig. Bass continuation lines disappearing

The issue was a consequence of the switch to fractions in 3.1: the ticks of a Fig. Bass element were written in the file as fraction but read back as an int, which in most cases resulted in a 0-duration element._

The following example is a Wikipedia of figured bass. At bars 48 and 66, there are perhaps multiple continuation lines, but they are exported incompletely into Musicxml. At 48, the xml only shows the end of extent, while in 66, there are only two marked as continue, no start and stop. This should be fixed.

Attachment Size
bsgbass.mscz 44.28 KB
Fix version