Today, I thought I would go through the complete process of composing and orchestrating a few measures of the first movement.
This chunk is from right before the climactic entrance of the theme, near the end of the movement. Originally, the theme was going to have D as it's tone center, but as I played around with some chord progressions to support it, I came up with one that had a continuously descending bass line that I really liked. The only problem was that it put the double basses down below the range of most basses. So I decided to transpose the whole section to F, which meant this transition had to modulate from D to F without being too obvious, the entire movement until this point had D as it's tonal center, and a keychange so close to the end wouldn't make any sense unless it came back to D. Which I won't. (More on that in another post.) So it has to be hidden enough that it doesn't jump out to the ear.
First, I played around with some ideas on the piano, and came up with this little motive I liked:
This would work well as a means to unify the section and make the transition more smooth. As it went on, I changed it a little, and figured out about 12 measures of it. The first two iterations of the idea above have an A sounded in the bass instruments right before them (where that quarter rest is, right before the notes). A is the note that pulls the most strongly to D, which is not where I want to go. So the next two iterations have a C in the bass. This pulls very strongly to F, but would be too obvious. So I started improvising.
This is a little rough, but you can hear how some ideas stuck and some were thrown out. As I played, I started to get a better feel for the way I wanted the harmonies to flow. After these two takes, I was pretty sure I liked the first few measures, so from then on I started with the next problem spot.
After this one, I was pretty sure of what the harmony would be for the next few measures, so I started from the next trouble spot. Generally, this is the way I improvise. I go through over and over, focusing on areas that are awkward or not what I want, until I begin to get some sense of a progression that works well.
You can hear how there are a few rhythmic and melodic ideas that carry over through all of these takes, even though some chords and ideas don't. Usually, these are the ones that make it onto my computer.
Here, I have pretty much nailed down the harmony, and am just fooling around with the last few measures before that really solid C chord, which leads to F and the melody.
At this point, I was ready to start writing all of this out. Now, when I go to transfer the ideas from my head (or iPad) to the notation software, I am still thinking critically about them, and so things change even as I write stuff down. The biggest thing I changed at first was rhythm. I wanted the rhythm to be exciting, and not too stable. So every time there was a repetition of melodic ideas, the rhythm would change.
That's the barebones fanfare. It needs embellishment, right? I started with the woodwinds and strings.
Next, the percussion, mostly mallets, doing rhythmic patterns and hits (like at the beginning of the movement; remember the last post?) along with harp and celesta, and some more moving stuff in the brass parts to make things exciting. In particular, trumpets 1 and 2 have a soaring line at the beginning, and trumpets 3 and 4 double the bells later on.
This fit together well, except that it was too cluttered. Every person was playing all the time. This doesn't make for a good thing in several ways. For one, there is no clarity of timbre. You can't tell whether the trumpets or the horns have the most interesting part, and the woodwinds just get on your nerves after a while. And this makes a piece lose momentum. For another, it makes the players frustrated, as they have to blend with eachother all the time, and often feel that their part is lost in the wash of sound. Also, it is very tiring on the ear to hear loud tutti sections where everyone plays everything.
So, I decided to cut out a few parts. First, I took the woodwinds out of the first measures where they had runs. I gave the runs to the horns instead. This did two things. First, it allowed the brass section to really assert itself before becoming part of the larger texture, and let the horns really show off their brilliant upper range. Second, it allowed me to hold the shrill woodwind parts in reserve for later, when there was more tension. This gave the section more momentum. Also, I took out the flowery trumpet lines that doubled the mallet parts. These were low in the trumpet range, and quite tricky, but wouldn't be heard anyway, since there was so much going on. This allowed me to split the brass section into brighter and darker instrument sounds (trumpets/trombones vs. horns), and mix and match them in a more interesting way. The result?
Well, there you go. A somewhat disjointed, unneccesarily verbose explanation of how I came up with the transition from D to F leading into the main theme. Next week, I'll talk about the form of the first movement as a whole, and how that relates to the entire symphony. Farewell until then!