Issues with dotted notes and adjacent noteheads (second)

Hi areapiano

For future reports, could you include the operating system and MuseScore commit?

The problem also exists in one voice... MuseScore 11c5868e00

Expected result

In this case where the two notes are dotted, MuseScore should move the dots up or down. The use can do it himself though by double clicking the dot and moving it so it looks like a minor annoyance.

See also #15434: Incorrect positioning of augmentation dots

I have a reasonably thorough fix in place.

https://github.com/MarcSabatella/MuseScore/tree/21870-dot

It handles seconds and also unisons, both in a single voice as well as between voices. It doesn't handle clusters of more than 2 notes the way Gould would prefer, but it's no worse than before, and these can still be adjusted manually if you like.

Two noteworthy aspects of the fix that I thought I'd give a heads up on before submitting a PR:

As part of the fix, I changed the existing dotPosition property of Note into userDotPosition, and added a new dotPosition property for internal use - the same basic model user for mirror / userMirror. This means the plugin framework now will access userDotPosition instead of dotPosition. I think I got that right, but am not sure.

Another existing behavior I changed is to honor UP and DOWN dot position even for notes on spaces. This is necessary to handle unisons correctly. There are potentially *three* dot positions for a note on a space: on the same space, the space above, and the space below. But I didn't want to add a new Direction to our enum and have to update everything, expose it in the UI, etc. Luckily, this was not necessary in order to handle the conflicts between unisons. Upstem voices are supposed to offset one dot up a space; downstem voices are supposed to offset one dot down. So I adopted the convention that for upstem voices, UP moves the dot to the space above; DOWN leaves it on the same space. For downstem voices, DOWN moves the dot to the space below; UP leaves it on the same space.

A couple of images from my vtests:

See the unisons between notes on spaces in the second example for what I mean about dot position for notes on spaces.

Marc, it's hard to tell from your tests, but this line still leaves me concerned: "Upstem voices are supposed to offset one dot up a space; downstem voices are supposed to offset one dot down."

In all the choral examples I've been perusing, downstem or upstem dotted notes on a staff/ledger line _always_ put the dot up, unless there's a collision which requires the dot to be exposed elsewhere. Your examples seem to show this happening correctly, but your explanation says otherwise.

Can someone enlighten me with some written rule about dot positions and lines?

Sorry I wasn't clear. I didn't mean *all* upstem notes need to have dots displaced upward. Just the specific case I was talking about - unisons between notes on a space, and only the cases where they cannot share a notehead. In other words, the unison treble clef A's in the second measure of second example. Similarly, the bass clef C's in that same measure are in a downstem voice, so one of the dots is forced *down* a space.

Accidental double post

Thanks for the clarification Marc. I could tell from the examples where dotted noteheads where shared that things seem OK, but didn't see anything showing the dot positions on unshared downstem dotted notes on lines. I just want to make sure things are OK. I await the PR and merge!

https://github.com/musescore/MuseScore/pull/797

I had written this up before but never hit "Save":

I should say that while I use the terms upstem and downstem here, it really isn't about whether the note/chord itself is upstem or downstem. It's about whether the *voice* of the chord is one that normally defaults upstem or downstem. A note on a line in voice 1, for example, always has the dot *above* the line when there is no conflict (ie, no second or unison), but a note on a line in voice 2 always has the dot *below* the line when there is no conflict. This much MuseScore was already getting right, and I didn't change how that works, so I didn't add any new tests for that case.

The code that I added deals only with seconds and unisons. And most such cases are covered by my tests above. You can see handling of seconds within a single voice, between two voices with upstem voice and on, between two voices with upstem note on bottom. And you can see handling of unisons in a single voice (both upstem and downstem versions), unisons between voices with heads shared, unisons between voices with heads not shared.

"but a note on a line in voice 2 always has the dot *below* the line when there is no conflict."

This is the line I disagree with. I have never, and I mean _never_ seen a downstem note with the dot below the line. I've gone over a lot of published score, esp multi-voice choral, and this simply never happens, unless a collision forces the dot elsewhere.

Measure 1 is the MS default for downstem dots, measure 2 is how I've always seen them.

You say you've never seen this, but were you actually looking for this specific situation? A quick perusal of scores on my shelves shows it is practically universal. All of the scores (from a variety of different publishers) do this consistently - conflict or no conflict. I'm not saying some publishers somewhere might have chosen to deviate from this, but I suspect if you look more closely, you'll discover the majority really do drop dots below the line for lower voices.

The go-to reference these days is Elaine Gould's "Behind Bars", page 56, and she is explicit about this as well. First sentence under Dotted Notes / Double stems / Lower part on a line", it says "Drop the dop into the space below the lower part", and shows an example that looks just like how MuseScore has always done it (and MuseScore has been around longer than Behind Bars, so this is not a case of MuseScore blindly following Gould).

BTW, I will say that of the scores I consulted, *one* did seem to use a rule where the dots on lower voices were dropped most of the time but not in one case there the two voices were a third apart - maybe to make this specific interval clearer?

I can confirm Marc's view, from a cursory look at a pretty recent song book ('Band- und Chorbuch zum Gotteslob', Strube Edition, 2013), dotted note in a line has dot up for voice 1/upstem note and dot down for voice 2/downstem note

Can someone scan the relevant pic from that book? I've only found one example so far for this, so I'm trying to see under what conditions this rule is used. This seems like more of an exception rule, rather than a default that should be applied.

For example, if I write a line in voice 2, and keep voice 1 empty, all dots will go down even though there nothing above the notes except rests. This is considered correct?

FWIW, it's also an area where Finale, Sibelius, and LilyPond all are all in agreement - lower voices all have dots *below* the line in all cases.

In fact, the more I look, the more universal this rule appears to be. I keep looking and have yet (in over a dozen tries) to find a single publisher who does *not* adhere to it - not as some sort of unusal exception, but in virtually all cases. In other words, aside from the one chord in the one piece I mentioned, *every single occurrence of a dot in a lower voice is below the line* in all the examples I have looked at (probably at least 20 by now, since you got me curious :-).

I suspect this will turn out to be one of those cases where you've been seeing something all your life and it's never really registered.

Now, the specific case where an upper voice happens to contain a rest - I could imagine some publishers might choose to go with dots above the line there. I wasn't looking for that case specifically. But a quick search for that shows at least the first example I found still has dots below the line for lower voices even if the top voice contains a rest.

I was actually hoping that I could see the scan of the page from Behind Bars, with any example she might present since I don't have the book.

Regarding the universality of this rule, I don't see it. After looking through many songs in my various compilation songbooks, and some choral scores from various decades, the majority _don't_ adhere to the rule. Some even conflict through a piece. Even a new choral score from SoundForth (distributed by Lorenz) dated 2012 doesn't obey this rule. If companies in this decade don't get the "rules" right, and surely they use software to layout music, who's right and who's wrong?

Well, I don't know that it's ever a question of "right" and "wrong". Mostly a matter of what is most commonly used and generally accepted/recommended versus what some individual publisher might happen to choose to do instead.

I'm really surprised you are not finding this rule as universally agreed upon as I am. Statistically speaking, this should be practical impossible - I mean, I consulted something like 40 different scores from at least 20 different publishers, and I finally managed to find *one* (and only one) that breaks this rule. The odds of this being coincidence are astronomical. Although FWIW, I am not limiting my consideration to choral music, as that is not the only or even most common situation where multiple voices are used. So I am also looking at piano music and guitar music. But even within the specific world of choral music; I'm still finding it virtually universal.

Anyhow, here is the relevant passage from Behind Bars:

Can you post a sample of two where you see this rule being violated? Also, are the publishers reputable ones? I tended to focus on large internationally-known publishers, on the assumption their engravers would be especially well-informed.

Here, FWIW, are links to the first relevant scores that show up on ISMLP when you ask for a random page. I skipped any that contained no multi-voice writing or that were "ancient" and used different style of notation altogether.

http://conquest.imslp.info/files/imglnks/usimg/a/a3/IMSLP242605-WIMA.cb… (see last measure, bottom staff)

http://sausage.whatbox.ca:15263/imglnks/usimg/0/06/IMSLP82867-PMLP14173… (see page 2; pretty much every measure contains an example of this)

http://sausage.whatbox.ca:15263/imglnks/usimg/5/56/IMSLP274850-PMLP4463… (see ms. 3, 8, 10, etc - but this particular score does show dots above line for bottom voice *if* the top voice is a full measure rest)

http://conquest.imslp.info/files/imglnks/usimg/f/f8/IMSLP114597-SIBLEY1… (see pg. 12, bottom staff, last ms. second-to-last system)

http://javanese.imslp.info/files/imglnks/usimg/7/73/IMSLP06175-Ravel_-_… (see very first measure)

http://sausage.whatbox.ca:15263/imglnks/usimg/2/23/IMSLP215119-SIBLEY18… (see organ part on pg. 4)

http://petrucci.mus.auth.gr/imglnks/usimg/4/4d/IMSLP213511-PMLP356033-f… (see pg. 3)

Finally, I found one that does not follow this rule:

http://javanese.imslp.info/files/imglnks/usimg/4/48/IMSLP60048-PMLP1229…

FWIW, Werner had some concerns with implementation details in my code so he made some changes and submitted a modified PR:

https://github.com/musescore/MuseScore/pull/808

As far as I know, the changes are internal only; there should not be any user-visible difference.

The PR is merged in 408b74df94. Thank you!

Automatically closed -- issue fixed for 2 weeks with no activity.