# Incorrect secondary beamings + Tuplet brackets always shown in beamed groups

After importing a score and resetting everything to its defaults, I was intrigued by one Dorico change of beaming, as I know how carefully it was designed in regard to notation conventions.

I’ve searched some scores, and found a very similar case to my own. This is from an Universal Edition score, followed by the Dorico’s default beaming (factory settings, except “Break beams at beat boundaries”):

For me, this is rhythmically misleading, or at the very least rhythmically much less clear. The first impression I have is that we’re dealing not with a 4/8 measure, but with 2/8 + 3/16 + 2/8 + 2/8.

I would guess that the Dorico team followed a recomended “precaution” found in Stone (pg. 111) and Gould (pg. 212), stating that tuplets should indeed have the secondary beams split, or have tuplet brackets to make them clear (although not necessarly both). The examples they present are eloquent. But:

1. they stress from the beggining that the beam structure must reflect the rhythmic structure (and don’t suggest that the tuplet beaming should overrule this);
2. it is stated that “The number of beams separating the groups is equal to the duration of the groups they separate” (Gould, 156);
3. they only present examples with tuplets that add up to the value of an eighth or more (not less, like in here), and where an “eight beam” always separates at least eight durations (again not like in here).

I conclude that the main purpose of breaking the secondary beam is to be more clearly distinguishable the tuplet notes from the non tuplet ones. In that sense, the tuplet bracket and/or the differentiation of the “32nd beams” from the “16th beams” seems clear enough to me.

Anyway, the important thing here is that rule in 2., which is the same as Dorico’s “Number of lines equal to the duration of the secondary group” (in Beam Grouping/Secondary Beam Groups options), and which Dorico doesn’t always follow (even when there’s no tuplets involved).

Bellow are other incorrect cases I found as I was fiddling with random rhythms (by incorrect I mean Dorico bypassing that general rule, and its own in the Notations Options), presented in both 2/4 and 4/8 metres (I also signaled in non red color other smaller inconsistencies, although not necessarily incorrect).
Besides tuplet cases, in the 4/8 metre versions Dorico also breaks (incorrectly) some secondary beamings to an “eigth beam” (I guess to separate more clearly the half of the beat).

Bar 5 top staff defies my understanding, specially to where the middle of both beats are. And if I select the option “Do not split secondary beams”, the results are even more incomprehensible: some beams still split (also the 64th tuplet beams, although now with a correct 32nd beam), and an other previously unsplit beam become split.

Dorico’s default beaming generally works quite well, and for particular cases it provides the tools to meet pretty much any requirements.
But it shouldn’t force us to correct things that, IMHO, goes against the general rule. I do think these excessive “precautions” – if conceived as such – are incorrect, to a point that the rhythm clarity goes in the opposite way it was aimed to.
Unless I missed some “hidden” setting (please point it out to me, if I did), I’d ask the team to either reevaluate this default behaviour, or the algorithms, or to add an alternative option for these cases.

Finally, there’s this small bracket issue:

On Engraving Options, I have set “Show tuplet brackets only when necessary”, but, unlike the UE top example, Dorico always shows them in beamed groups (even when there aren’t any tuplet rests).
As one can “Hide ratio and bracket” on repeated tuplets – an option which isn’t bypassed in this case (and if set would contradict the previous bypassing “precaution”) –, I think that there should be also an added option to disable brackets (but not the number) in beamed groups.

Thank you for considering this.

secondary beaming.zip (377.5 KB)

2 Likes

I tried recopying the first example above (looks like SCORE) and was able to match almost everything, only by making the following edits:

• Despite having already set {Notation Options > Secondary Beam Groups > Number of lines in primary beam within secondary beam groups → Number of lines equal to the duration of the secondary group} I still had to manually beam together all the notes of each quarter of the bar (eighth duration) and re-break the 32nd beam in each pair of triplets. It would seem the setting is not doing what it says, but if I am misunderstanding, please explain.
(By the way, the illustration of this setting in the dialog could be more informative if it included some other note values as well as 64ths in the pictures.)
• Manually hiding the brackets on all tuplets but the first one
• Of course Notation Options > Accidentals > Basic > Second Viennese School
OT – More required edits, for the sake of thoroughness:
• Since the first accent is outside the slur and the last one inside, one of them has to be hand-tweaked. I decided the make the first one the exception, since that slur also needed manual position adjustment. In order to stop the accent and slur fighting for position, I flipped the accent above and dragged it all the way back down in engrave mode. Then repositioning the slur was fairly easy, but I had to redo it several times as I tweaked other things.
• The starts and ends of the hairpins in the source are rhythmically ambiguous. To match these more closely I temporarily made 64th notes in order to place 3 of the 4 of endpoints between the note positions.

Yes, I think it’s exactly that. For the brackets “issue”, the “Show only when necessary” setting works only if I unbeam the tuplet groups.

And sure, one can luckily tweak almost anything as one pleases. I too did that. But correcting the beaming alone of that last beat with the tied note is quite cumbersome. The passage in question is from a 165 page score (Boulez’s Dérive 2) full of similar tuplet cases in all instruments. Can you imagine doing all of that on a regular basis?