Beams in 1.1

Tripple beams are still unpredictable. Check out the examples below.

I’m not sure exactly what you consider to be the problem here: in order to avoid wedges, if you use the regular, non-widened distance between beams, Dorico snaps them to a space; you can use shallower slants if the gap between beams is widened, as your third picture shows.

Sorry, I should have been more explicit.

2.png - the settings indicate that the narrow beam should be 1/8 space, it is not.

3.png - the settings indicate that the slant of only two notes should be 1/4 space, it is not.

4.png - the settings indicate that the slant of only two notes should be flat. it is not.

Daniel, did you have a chance to look into it?

Please don’t bump threads! I have other duties to attend to in addition to being here on the forum, so if something takes a bit of time to look into, I can’t always drop everything else I’m doing to immediately provide an answer. But I keep every thread that requires further attention open in my browser and when I can make the time I will look into it and provide you with an answer. Sorry for the delay: I will come back to you as soon as I can.

Really all of the slant values in Engraving Options should be considered as ideal slants: they are not necessarily the slants you will end up with once the vertical snapping algorithms (designed to produce legal sit-straddle-hang positions and to avoid wedges with staff lines) have been occupied. These are subtle, and involve dimensions like ensuring that the inner beam doesn’t get positioned within half a space of the outermost staff line.

As I said in my first reply:

Again, the slants in Engraving Options are the ideal slants, but Dorico prioritises the avoidance of wedges and the production of legal sit-straddle-hang positions over the ideal slants.

It looks like something is not quite working here: any value greater than 0 will work as an ideal slant, but an ideal slant of 0 is not what the program expects. I don’t consider this an urgent problem, since I think it is highly unlikely that you would actually want an ideal slant of 0 for all two-note beams.


I attach a more comprehensive example, with all possible beam angle settings set to 0.
It doesn’t really matter, if it’s a “0”, “1/16” or “1/8” – the result is unexpected
and there’s no way to achieve cohesive shallow slants across the whole score.

I think the approach used by the Patterson Beams plug-in for Finale is much more predictable.
I’ve never had a problem with inconsistencies. All beam slant values are considered as “maximum allowed angle”.
It’s just much more practical and intuitive.
Screen Shot 2017-07-23 at 1.16.10 PM.png

I’m sorry you’re not happy with what Dorico does. However, it is not really intended that you should set any of these values to 0. Perhaps I should modify the dialog such that you cannot set a slant of smaller than a quarter of a space for any of these values, since that is really the practical limitation. Even when the values are set to 0, Dorico will still want to assign beam slants to rising and falling figures, and the other phases of beam processing will snap the beams to legal positions based purely on the ideal stem length, and that will result in beam slants in some cases.

My point is, it would be more predictable if a setting indicated a maximum slant not an ideal slant.
IOW, the beams should only snap to a flatter position, and never exceed what’s in the settings.

Daniel, I know this is how it’s designed in Dorico, but I’m not sure I understand the reason behind it, and why you think it’s better this way.

and BTW, I love Dorico. It produces the best note spacing out there, finally looking like professionally engraved.
And I am your avid supporter here in Hollywood. I just think that the way Dorico handles beam placement is not as good as in Finale with Patterson Beams.

I think I’ve explained before that beam placement is processed in three phases in Dorico: first, we determine whether the beam should be flat, slant upwards, or slant downwards; then we determine the ideal slant; then we snap both ends of the beam to legal vertical positions. The reason that we do not express the beam slants in terms of a maximum is that we decided that coming up with legal beam positions, ideally those that avoid wedges, is more important than producing a slant no larger than that specified in Engraving Options.

If there are specific (non-contrived, real musical) situations in which you are experiencing problems with reasonable settings on the Beams and Notes pages of Engraving Options, then please let me know what they are.

Not sure why would you consider my music to be a non-real musical example, but ok…
What I am trying to achieve is Henle style beams, where most of them never cross spaces.
In my example, regardless of the settings (I chose zero, just to make it clear), the group of two 32-notes in bar 29 and 33 will always have a slant of 1 space. In this particular case, the two notes are only a second apart (1/2 space angle), and yet the beam is steeper than the general direction of the noteheads. It looks plain wrong by any classical standard.
The algorithm shouldn’t be allowed to increase the beam angle. Stem length should be adjusted instead.

I didn’t mean to imply that your music was not a real musical situation: it was a clumsy way of saying that I didn’t think setting all of the desired slants to 0 was a musical approach. Please accept my apologies for any offence caused.

I need to talk to James, our beaming guru, about whether or not it is possible with the current set of settings to ensure that 32nd beams and shorter can follow the same 1/4 or 1/2 space slants that Henle would use; I’m not sure that it is, though I agree that it should be. We’ll talk about it here and I’ll come back to you when I can.

Thank you Daniel.

Let me bring this old topic back just to contribute with real published scores of different beam angles examples. (BTW, I think I did talk about this some time ago in this forum, but I couldn’t find the post.)
I’ve also read your huge analysis on the subject in the Dorico Development diary, part 10.

In general, I like Dorico’s default settings for beam angles, although there are some specific situations where I would prefer more control over the results.
If we take a look at G. Henle’s use of beam angles, most of them have slight angles and are practically flat. If we try to reproduce this behavior in Dorico, some cases still get too steep angles, as in the attached image.

I know that 16th beams, 32nd beams and so end up with a large angle as it’s the only slant that does not produce wedges on at least one of the beam lines, as Daniel has said in other posts, but it would be awesome to have an option to let Dorico take into account secondary beams to align them as in the desired examples below, as it looks a lot cleaner and tidier.

In the first image, there’s only one case (bar 6) in which Dorico gets, in my opinion, a nice beam angle that does not produce an ‘interference’ with any barline. The second one is just for the sake of comparing extreme cases.
[Green: desired, Red: not wanted]


The real cause for the greater slant in these cases is Dorico’s desire to avoid a beam being positioned within half a space of the outermost staff line, and we do plan to relax this restriction, which will allow you to obtain the 1/4 space slant you’re looking for.

1 Like

This is really good news, Daniel. For some time, Dorico’s beam angles in specific cases have bothered me but I haven’t had the time to compile a list of examples. Thanks to both Albert and Sugar for their examples which illustrate the point well. The irony is that in an attempt to avoid wedges, Dorico often actually produces them, as in Albert’s examples. In the beams_henle example, I would even say that Dorico’s beam angle for the first group of LH 16th is irrational. Why should the inner beam of the first F have to start on the second staff line? The inner beam of the F at the beginning of the second group starts on the bottom line. Henle’s beam angles produce a consistent result and, as Sugar rightfully said, very much like Finale’s beam angles with the Patterson beams plugin. A beam connecting two notes which are a second apart shouldn’t have an angle of an entire space. Because Dorico does this inconsistently, it disturbs [sight-] reading, as beam angles are one of the elements one uses when reading music. Seeing an interval joined by a shallow beam angle followed the same interval at a different staff position but joined by a steep beam angle is visually confusing. FWIW, Gould’s examples in pages 19-21 would seem to support this.

That would be super! This is something that I still have to do by hand and that there was no way to scale up.

Continued development of beam angles is great news. Would love to see Henle-style options.