The idea of your music: the music model
Dorico’s innovation with the most far-reaching consequences is its thorough approach to modeling musical concepts and their intricate interactions. This might not be apparent right away, as it is not so much a single feature but the accumulation of many small ways of capturing the meaning of notation conventions and not just their appearance.
An example case: the voices paradigm
Have a quick look at Finale’s documentation for the Layers feature (the equivalent to Sibelius’s voices).
In condensed form (italics added for emphasis):
- “Each staff in Finale has four transparent layers of music.”
- “[…] each layer may be taught to flip its stems […] to help distinguish the multiple voices.”
- “In general, you’ll want the stems of Layer 1 to flip up, but only when Layer 2 is present […]”
- “Furthermore, you’ll probably want ties to flip […]”
- “Therefore, you’ll probably want to select options as follows.”
- “The settings for Layers 3 and 4 are up to you, since [settings] probably depend on the piece you’re notating.”
- “In addition, you may wish to specify that the placement of rests […]”
- “You tell Finale how far out of the way you want these rests to appear […]”
- “In the usual situation, you’d enter a positive number […]”
In summation: while Finale offers a solid technical framework and all necessary tools to create the graphically correct appearance for any complex notation, there is an implicit assumption that users can not expect the software itself to infer how to reasonably use these tools.
It turns out that this can indeed be expected. To get there, Dorico does not simply improve upon the status quo, but goes back to first principles. For voices, this means breaking with a presumption taken mostly for granted today: the four-voices-per-staff model.
Sibelius works this way (voices 1 and 3 are up-stemmed, voices 2 and 4 down-stemmed), as does Finale (see above) and even LilyPond — which does support more than four voices in a somewhat straightforward way — does acknowledge the four-voice model in its predefined commands.
Dorico, however, does not make any assumptions in this regard. Instead, it just requires from the user one fundamental decision before starting to input notes: is this voice supposed to be up-stemmed or down-stemmed?
The caret appearing when note entry is started (Shift-N) in Write mode will show a small stylized note symbol, up-stemmed by default.
By pressing Shift-V the user can call for a new voice to be created, which will — nominally — be down-stemmed. (Similar to the behavior of Sibelius and Finale, these directions will take effect once there are two or more voices present; one voice alone will have its stem directions determined by the usual rules.)
Pressing Shift-V again prepares a new voice, with stem direction reversed once more. I use the word “prepare”, because a voice is only ever actually created if notes are entered.
This is where the old paradigm is abandoned. Instead of choosing the “correct” voices from a limited set, users create exactly the voice constellation that is needed; for example: one down-stemmed voice and five up-stemmed ones. Once voices have been created, users can easily toggle through them by pressing V, allowing for polyphonic notat… Wait, wait, wait – just a minute. What? One down-stemmed and five up-stemmed? Who would ever need such a weird voice construction?
Well, one Ludwig van Beethoven would:
In all fairness, this short passage from the third movement of op. 27/2 does only really need five voices. I opted to use a sixth one on the chords, for clarity — and Dorico let me.
If you try to create this example in Sibelius (which you will fail to do without heavy workarounds, since you are at least one voice short), you will spend some time hiding a lot of excess rests, created because Sibelius insists on each voice being complete within a bar, also meaning that a particular voice can technically only start or end at a barline.
Dorico does not offer a feature to hide notational elements yet, so where are these rests? Again, Dorico’s voices work opposite to the way we are accustomed; they can start and end literally anywhere. Appropriate rests shown before or after the extent of a voice are added by Dorico automatically, unless users opt against it.
Instead of having four determined voice “tracks” which can be used to mimic the concept of independent voices, the actual concept is deeply implemented in Dorico. And instead of thinking about where to best place a voice in the available technical framework (and thereby necessarily thinking about that framework itself), users only need to concern themselves with a genuine musician’s decision: up-stemmed or down-stemmed?
Dorico’s promotional material claims that there is no limit on the number of voices for a stave, and I have not been able to create an example to refute this. It surely is an impressive achievement in algorithm design, even though there is, in practice, rarely a use for more than three voices, let alone an infinite number.
But okay: here is the first full entry of all 40 voices from Thomas Tallis’s Spem in alium. I hope we can agree that this example is exhilarating, but not really helpful for anything (click for PDF).
So rigorously from the ground-up is the program designed that finding one’s way into working with Dorico often serves as a useful refresher course in notational thinking in general.
Here are the articulations that Dorico offers its users during basic note entry: To the right you have your staccato and tenuto, invoked by the shortcuts ] and #. Adding the Shift key will give you the respective shorter versions of staccatissimo and staccato-tenuto. Likewise, [ will produce an accent, or a marcato with Shift. Convenient, so far.
But what are these two symbols in the bottom left corner? There is a good chance that a number of readers see them for the first time – that is, unless they are also students of prosody, where similar symbols are used to notate patterns of stressed and unstressed syllables. Arnold Schoenberg introduced them to music notation with the intention of clearly showing how a rhythm is supposed to be articulated against an established overall meter. It would be interesting to hear in the comments from those who use these two articulations on a regular basis, but I won’t hold my breath.
Given the relative obscurity of the “stressed” and “unstressed” articulations, one might easily object their inclusion here. After all, aren’t there many types of articulation that are much more widely used? Where are the circle, the cross, up-bow and down-bow?
But think about it for a minute. The one thing Dorico’s eight articulations have in common, and which distinguishes them from any other “articulations” (as we may be accustomed to think about them), is that their meaning will not change with context.
Put another way, a circle denotes a harmonic on a string instrument, but on a horn it means an open note (after a muted passage). The cross (or plus) will stand for left-hand pizzicato in strings, but for hand-stopping in horns. Up-bow and down-bow really do not make sense anywhere except within a string part. It takes a moment of adjusting one’s preconceptions to realize that both Sibelius and Finale (and most other scorewriters out there) consider symbols as articulations which, on closer inspection, signify playing techniques.
Actually, to say that they “consider symbols as articulations” may be even a stretch for Dorico’s competitors. It would be more accurate to say that they consider notes, and then also symbols that can be stacked above or below these notes.
Case in point: the image to the right shows something that passes as a valid notation in Sibelius. Dorico, instead of just placing glyphs, makes a distinction between notations that have a universally musical meaning (the actual articulations from the left-hand panel) and those which carry technical information (found in Write mode’s right-hand panel, under Playing Techniques).
The differentiation does not stop there. Articulations themselves are separated into three categories.
To use Dorico’s terminology:
- Articulations of duration (the articulation panel’s right half);
- Articulations of force (accent and marcato);
- Articulations of stress (the Schoenberg symbols).
Only one from each category can be applied to a note or chord at a given point. Once one takes articulations seriously as musical concepts and not just as symbols to be placed in relation to a note, this makes good sense. While it is possible (not in Dorico, though) to have a note be simultaneously marked as “stressed” and “unstressed” on the notational level, it really can only be played (articulated) either the one way or the other.
This sophisticated understanding of articulations also is the key to another detail that might look odd at first. Dorico provides a dedicated staccato-tenuto articulation. Visually it is, of course, a combination of the two glyphs for staccato and tenuto. Musically, however, a staccato-tenuto is only superficially a combination of staccato and tenuto.
One and the same note can not be deliberately shortened and at the same time deliberately held out for its nominal duration, of course. An actual staccato-tenuto is a two-step procedure of first shortening a nominal note value and to then taking care to actually emphasize this shorter virtual duration (or the other way round, if you like).
Dorico wisely acknowledges this, even though technically it is not necessary for producing the appearance of the proper notation. In Sibelius, a staccato-tenuto is created by stacking line and dot. Finale, in contrast, not only allows to have the kind of hit-and-miss implementation available in Sibelius, but to additionally add the line-dot-combination glyph, making it possible to mark up a note with a stack of two lines and two dots in total.
Back to first principles: dynamics are dynamics; ottava lines are…clefs?
It is obvious that one guideline that has been observed closely by Dorico’s developers is to not go for solutions which merely give a straightforward technical way to a specific notation’s appearance.
In other programs, a dynamic marking is either a symbol or a text object or a line, depending on what is easiest to do with the existing framework. For Dorico, it is neither; instead it foremost is simply the concept of a dynamic. This concept can then be expressed by the user in different ways, be it a symbol, a text or a line — even combinations. But underneath, it always is the same fundamental type of event.
In going to these conceptual lengths, Dorico shines where existing software starts to stumble. It is inconvenient enough that Sibelius users have to construct text-chain instructions of the “de – cre – scen – do po – co a po – co” variety by arranging individual text objects. But then they might have to do the same work again because of horizontal spacing in a part diverging heavily from the score. And the real frustration starts when such a carefully fine-tuned passage has unexpectedly to be changed substantially.
In Dorico, this poses no such problem.
Sometimes, following Dorico’s music model leads to elucidating surprises. If you thought that ottava lines are… well, lines – consider this: their effect is that any notes to which they apply are displaced to a different stave position. Dorico therefore has them share a category with the only other notational elements with a similar function; ottava lines are now basically clefs.
Since Dorico comes with its own fledgling DAW, having ottava lines behaving like clefs has also the clear benefit of keeping the pitch information of a score perfectly intact at all times. This, once again, is the opposite of the commonly used method of having the worst of both worlds: combining notes that refer to the wrong octave with a line that clumsily counteracts this by adjusting playback.
The meaning of the music
Dorico’s model attempts to comprehensively cover not just appearances, but meaning. It does succeed to an unprecedented extent. But music — being art and therefore an act of resilience — is not inclined to take this without a fight.
While the articulation system described above is perfectly logical, appropriate and well researched, the actual practice of seminal (and not-so-seminal) works might not necessarily care. The newest of the many reasons to lament György Ligeti’s passing is that we will never witness how Daniel Spreadbury explains to him that it is wrong to stack three “articulations of force”.
Likewise, there has never before been a scoring program that actually understands how to reasonably handle fermatas. Still, Richard Wagner managed to come up with fermata use cases like this one, which is as invalid as it is ingeniously apt:
It will be interesting to see how the Dorico developers choose to engage this eternal dilemma. Will they be able to expand their model without sacrificing its elegance and consistency? Or will we eventually just barter our Sibelius and Finale workarounds for ones idiosyncratic to Dorico?