The title is sarcasm. Yes, it is.
The Ukiah Symphony is planning to perform my piano concerto next season, and before that can happen, I have to get them a full score with parts. This has been a bundle of joy.
Here are a few examples of the kind of things I’ve had to figure out how to do in Finale (the software I’m using to score the music):
Bleaugh! Of course, it looks gorgeous NOW.
So what does this process of scoring music look like?
The process starts with the Cuebase sequencer file I used to produce the CD. The first task is to quantize the sequence. You see, when I want a staccatto note in Cuebase, it looks something like this:
If this came from a live performance on a keyboard, it isn’t nearly so regular. If I try to turn this into a score, I end up with a lot of very short notes, and lots of strange rests in between them. Editing those in Finale is actually worse than a root canal. On a plane. To Australia. So to make it easy, I want to convert it to something like this:
Lining everything up with the grid lines is called “quantizing.” There’s some other prep I can do, like separating out the different piano lines. That takes a little creativity, because it’s usually not as simple as left-hand/right-hand, or even bass-clef/treble-clef. But after going through all the tracks and quantizing everything, separating everything, and testing it to make sure I didn’t move notes around to strange places, I can export the Cuebase data as a MIDI file — which is kind of like the basic CSV common file format for musical notes.
Then I need to import the MIDI file into Finale, and it converts all those dashes into notes.
I have a lot of complaints about Finale, but I have to step back and take my hat off to them on this: they do an impressive job of converting those dashes into notes. But now the real work starts, because impressive is still a long way from adequate.
The first pass through the score involves putting in the key signatures. It helps if I put that into the Cuebase file, but the key signature doesn’t make any difference to the sequencer, and I always forget. Finale has some algorithms for guessing, but they are … well, a little bizarre. Which I suppose is understandable.
The whole point of key signatures is that the Western chromatic scale over an octave has 11 distinct notes in it, but the musical staff notation used since A. Nony Mous scribbled out the first madrigal has room for only seven notes over an octave. They make up the difference using “accidentals” — flats and sharps — that push you up or down from one of the seven notes. Of course seven goes into eleven roughly 1.57142857142857 times, so there’s some black magic involving something called “modes” — there are seven traditional modes — two of which are “major” and “minor.” And that’s the easy part.
The thing is, musicians spend about twenty years learning to read music well, and so they get understandably irritated when you break the rules they’ve learned and hand them something that they can’t read and perform easily. So you pick a key signature, and the basic idea is to try to minimize the number of flats and sharps you have to throw around in the music. If you’ve written something in C# minor, and then rewrite it in C minor, the only place sharps or flats show up is right at the beginning of the piece, where you specify the key signature. All the rest is accident-free, and the notes will be otherwise identical.
It’s equally easy to add the key signature to Cuebase or Finale, which is why I generally don’t go back to Cuebase and just start over — which would also be easy, since all I’ve done with Finale up to this point is push a few buttons.
The next step is going through the score, staff by staff, fixing the accidentals that remain. As it turns out, that 11 versus 7 black magic ends up with as many as four different ways to represent the same note — as a sharp, a flat, a double-sharp, or a double-flat. This does turn out to be useful, and it’s way easier to show why than to describe:
These are both identical passages, but the one on the right shows the flow of the music a lot better than the one on the left. You want to follow the natural flow, because it will make it much easier for the musicians to play while they’re reading it. And that’s something you want to do — make it easy for them to read. After all, if you piss them off, they don’t have to play your music well.
Unfortunately, Finale doesn’t always do a good job of deciding which way to write the notes — it chose the one on the left. So I have to go through and correct all of these so that they make sense.
As I’m going through this, note-by-note, I’m also correcting all the errors that arise from Finale trying to guess note durations. Finale does a pretty good job with triplets. That 21-note run up at the top? Not a whisper of a prayer. And, at the same time, I can correct the clef notation, to try to keep the notes more-or-less grouped in the middle of the five-line staff.
All of these edits are judgment calls. The music is already correct, if unreadable. I’m trying to improve the readability, and that’s ultimately a matter of opinion.
Once I’ve gotten through that pass, comes the sniff test — that’s the test you give a jug of milk in the refrigerator that has been in there since you aren’t-quite-sure-when. Fortunately, Finale comes with its own MIDI performance software, so I can tell it to play the music for me. I can listen for errors.
It sounds terrible, by the way. Standard notation is intended to facilitate the performance of a human being who spent twenty years learning how to read music and play it well. The truth is, there simply isn’t enough information in the score for a computer to figure out how to play it. But — and this is the important part — listening will at least tell you if you have the right notes, with the right rhythms.
Now the tedium begins. Every note has potentially a dozen or so “articulations.” Staccatto. Marcato. Tenuto. Tremolo. Pizzicato. Palmetto. Con Gigolo. Mio Spaghettio. Yes, I’ve started making these up-o. Made-you-look-o.
Remember all those short notes I “quantized” way up at the top? Now I have to recover that lost information by putting a little dot over each note that should be played short. More decisions: staccatto, spiccato, or marcato? They’re all kind of the same — but they’re all different, with different markings. Each instrument will perform them differently, and each musician will interpret them differently. Get a little alcohol in them, and they’ll fight about it.
Then there’s phrasing — slur marks to indicate phrases. This is a tough call, because a lot of phrasing is just part of that twenty-year learning curve. Throw a bunch of dots out there, and musicians will make phrases out of them, and usually the right ones. All you’re doing is giving them hints, and there’s no point in insulting them with the obvious.
Even worse, slurs are also used to indicate bowing in the strings, and breathing for the wind instruments. You don’t want your oboist to pass out because you created a single long phrase of twenty-five measures. Of course, they won’t do that, but then you might as well not waste the ink on a phrase mark that is going to be ignored anyway. I always find myself swinging between minimalism — hey, they’re musicians, let them figure it out — and the kind of obsessive notation that I have to put into the Cuebase sequence to get it to sound right.
Dynamics. Forte. Mezzo-forte. Sforzando. Subito piano. Crescendo. Diminuendo. I absolutely hate this part, because it’s like trying to pick your nose with a hammer. Dynamic markings are a very blunt instrument. There are only eight gradations from inaudible to deafening. Fine, ten if you want to include pppp and ffff, neither of which is playable. In fact, it’s arguable whether ppp and fff make a lot of sense — really, there are only six. Again, these are hints, but they can be extremely important hints — like when you want a horn swell that sends all the squirrels in the rafters running for cover, versus the sound of a dream falling into a feather bed.
Then there is the expressive text, like molto agitato. Or perhaps like butterflies hovering over a glass of pinot. I’ve seen very silly notations, like a passage in Wagner for the strings marked interminably, oh dear God will this ever end?
I usually end up putting in the tempo markings last, for some reason. That one is pretty easy: the Cuebase file has utterly precise tempos — it needs to. So I could add notation that says, slow from 111 bpm to 93 bpm. I don’t, of course. It would instead be Allegro (quarter = 110) followed by rit. The conductor is only going to take it as a suggestion, anyway.
Sprinkle in fermatas, indicating that a note should be held until the winds are blue in the face. Add a break, or even a Grand Pause, where there is pregnant (or perhaps blessed) silence for a moment.
Finally, there’s all the print niceties: measure numbers, page numbers, rehearsal marks, title, composer, copyright, etc., etc., etc.
Still not done, however. Now the parts have to be broken out. Musicians don’t read scores. They read parts. A flute part. A violin part. A piano part.
Here, Finale is truly your friend. It’s pretty much few button presses, and you have all the parts broken out for you. They still have to be individually edited, because sometimes Finale makes some rather silly decisions. In particular, it always seems to put the last measure of the piece on a page by itself, a single measure across the whole page with one stupid note in it.
And then — finally — let it sit for a week, and then edit it again with fresh eyes.
I’ve just gotten to the tedium on the second movement of the concerto. First movement is done and mellowing for that final edit. Third movement should be quick and relatively easy.