Things I Learned from Ligeti (a little long)

I’ve spent the last few weeks inputting two Ligeti piano etudes into Dorico. I’m patiently waiting for three-staff piano support before doing some more. Thanks to Claude, Rob and Fratveno for their help in fixing some of the issues I couldn’t handle. In the meantime, I’ve run into a few small problems/requests.

Time Signatures

  1. Ligeti places his time sig in the middle between the piano staves. I do not believe this is possible in Dorico.
  2. In Etude 5, the time sig is 3/4 (2/dotted quarter). Is it possible to mix the two styles of denominators in one time sig, or is it planned?


  1. Some articulations interact very oddly with slurs, and end up very high off the stem. I was able to override it by moving the articulation to the notehead side, and then dragging it in Engrave, but it seems a little buggy.
  2. I got into a habit of forcing all articulations in one voice to the stem side, before entering the second voice, because I could do that by selecting the entire bar. However, that forced placement didn’t ‘stick’. Once I entered the second voice, the articulations all flipped back to their default positions. I believe that once forced, articulation placement should stay unless overridden.
  3. Is there a way to parenthesize articulations?
  4. Currently the properties panel works on a ‘highest common denominator’ philosophy. If I select 8 notes, 7 of which have an articulation and one doesn’t, the articulation option doesn’t appear. Same if I select notes and a slur. It slows things down considerably, and it’d be nice if it were a little more forgiving and show all relevant properties for the selection.


  1. The (very verbose) dynamic marking in Etude 5 includes the word ‘molto’ somewhere. Dorico grabs that word and moves it to before the ‘p’ symbol, removing it from the text string. It’d be nice if it were able to distinguish ‘p molto’ from ‘p dolce, sempre legato, molto espresivo’ and not mess with the latter.
  2. I’m sure more flexible text editing is in the pipeline. Ligeti has local changes in tempo that are in a different style (lower-case, italic, non-bold) than the main tempo indication. I was able to fake them with shift-X text, but they still seem different than the text attached to rall. and allag. lines. I made sure to pick the same font and same size, yet something seems off.


  1. Is there (or will there be) an option to have cresc. and dim. lines that cross systems to reiterate cresc. in parenthesis at the start of each system, much like octave lines?


  1. I followed the original layout, mostly because I’d have gotten completely lost in Etude 1. However, adding octave or pedal lines broke the layout and I had to redo it every time I added them.

Special Requests

  1. An option for barlines that extend half-way between the staves for poly-metric music would be nice.
  2. It’d also be nice to be able to beam on the fly in open meter, maybe similar to the way slurs are created. One would enter an eighth-note, hit the key for ‘start beam’ and then all subsequent notes would be beamed until I hit the key for ‘end beam’. (510 KB)

It’s awesome you did this! Awaiting the team’s response

Yes, well done. would be nice not to have to fake the time signatures but be able to use the 1+1+1 for the dashed barlines but an option to show the sum 3/4.


Fantastic! Congratulations!

Just a little warning though that if you have plans inputting L’Escalier du Diable you will run into problems as the etude calls for no less than eight forte-signs and Dorico, alas, can handle no more than six :unamused:
L'Escalier du Diable.jpg

You’re right, it’s not. When we come to work on “big” time signatures for commercial music scores etc. we can look at introducing this possibility then.

No, it’s not possible. At the moment there’s only a single set of properties for each time signature in an alternating, aggregate or interchangeable time signature. We may be able to extend this in future.

If you have a simple reproducible case of this problem, please post it so we can try to fix it.

The property will stay set, but don’t forget that the placement is relative to the current stem direction of the note, i.e. you are setting it to be stem side or notehead side rather than above or below, so when another voice is introduced, it’s possible for that placement to result in a different position. I would suggest you only fuss with the placements once both voices are in place. We will hopefully have filtering operations soon that will make it easier to select e.g. notes in particular voices etc.

No, not at present.

We’re unlikely to change this. The Properties panel carefully works out which properties it can show based on the exact contents of the selection. We don’t want to make it possible to set e.g. articulation properties on notes that don’t have articulations, as that would potentially lead to unexpected behaviour later on (e.g. you add an articulation to a note later, and you suddenly find that it has got a forced placement). We do have a number of guards in place to try to prevent such data being kept around, but articulations in particular are problematic in that regard because they can be set individually on each notehead, and the program needs to resolve how they will appear. In summary, the current system is complicated enough that I would not be in favour of making it any more complicated.

Using a semantic representation for dynamics is convenient in some ways, but inconvenient in others. We will continue to think about how best to solve a situation like this in future.

You probably discovered that you can change the appearance of gradual tempo changes by editing the gradual tempo font style in Engrave > Font Styles, but of course this doesn’t automatically change the placement of such gradual tempo instructions. This kind of style is also often used in choral music to show the tempo changes on each singer’s line. I’m sure that in future we’ll be able to address this in a more semantic fashion.

It’s not possible at present, but I will make a note of it.

I don’t know what you mean by this: can you be more specific?

When we come to work more on polymetric music perhaps we’ll be able to take a look at this.

Yes, we agree that it would be nice to be able to set up beaming during note input, and this is on our wish list for the future.

Thanks for the detailed response, Daniel.

Re: layout - what I meant to say is that adding octave lines removed system and frame breaks that I’d put in previously and they had to be recreated.

Re: articulation placement - Etude 5, measure 6 shows such an example. The accent I’ve circled in red in the attached screenshot actually belongs to the highlighted Bb below it. The screenshot shows Dorico’s default placement of the accent.
Screen Shot 2017-03-09 at 9.30.35 AM.png

I found a few very minor errors in vilnai’s excellent engraving of the Ligeti, so here is a revised version. This makes me wonder - is there anyone here on the list who was involved in the engraving for Schott? I’d love to learn more about the original engraving, which is beautiful.
ligeti-1& (590 KB)

Sorry, I found some more stuff that I messed up - catching courtesy accidentals with all the added time signatures is difficult! But I think this version is pretty close. The main thing I’m missing is the ability to put accents in parentheses.
ligeti-1& (591 KB)