Thursday, March 28, 2013

Ricercare for Orchestra, v2.0 (Final)

It is finished!

Well, all but the orchestration, that is. I'm currently studying my way through Wallace Berry's Form in Music, and when I'm trough that (again) I'll pick up my copy of Samuel Adler's The Study of Orchestration and get with it.

I have some idea of where I'm going with the orchestration, of course, but it's been a while and my chops are kind of rusty in that area. I'll have, in addition to the strings, pairs of flutes, oboes and clarinets, bassoon and contrabassoon, then trumpet, horn, trombone and tuba. What I don't know is what to put in the highest octave above the trumpet. I'm thinking of a soprano saxophone, but I may even exceed that range. We'll see.

There was briefly a v1.1 of this, but it was a minor adjustment which I'll point out when we get there. Here's today's MIDI to MP3 conversion: Ricercare for Orchestra, v2.0 (Final)

The five-measure subject I composed as a five-part canon at the octave and single measures of delay, so the fifth measure of the canon has all five measures of the subject simultaneously. Note that there is nothing but stepwise movement in the subject with the exception of the signature leaps of a minor sixth and a diminished seventh at the climax of the head figure. Also, eleven of the twelve pitch classes are present, with only the minor second degree - a phrygian effect - missing.

Due to the chromatic nature of the subject, the answer must be real an not tonal, so we get what amounts to a direct modulation to the dominant region at measure six. Note that the counter-answer from measure eight on is the augmented form of the head figure, and it is also all stepwise movement except for the characteristic leaps starting in measure ten.

Then the counter-answer makes the expected leap of a diminished seventh - the c-natural in the viola part - into the next statement of the subject. This works because the minor sixth degree from the perspective of the dominant region is the minor third degree from the perspective of the tonic. Despite this elegant dovetail device, the re-modulation back home still has the effect of a direct modulation.

It is not possible to have another statement of the answer at measure sixteen, as countersubject one - the second violin part - also has the augmented head figure from measure thirteen on, and the required leap of the diminished seventh into sixteen makes that impossible. That's because the minor sixth degree from the perspective of the tonic is the minor second degree from the perspective of the dominant. So, I had to put this four measure episode in to reorient the voices for the next statement of the answer. This actually increases the interest of the exposition by avoiding the predictable monotony of five thematic statements in an uninterrupted row.

Note that the bass line in this episode finally gets some melodic leaps happening. It's a really cool bass melody, if I do say so myself.

At twenty we get another direct modulation to the dominant for the real answer in the highest octave of the orchestra. I'll have the flutes doubling the strings here, as they are just about the only instruments that can reach the high c-natural in measure twenty-one (A piccolo might even be better). This is the pitch climax of the piece, by the way, which sets up the drama as a 5-4-3-2-1 line, from the Schenkerian perspective. Note that every movement in every voice is stepwise, again excepting the characteristic leaps in the head figure, both in the original version, and the augmented version in the cellos. This is the smoothest fugue I've ever written, and I'm actually harkening back to the style of Palestrina here, except that it's tonal and not modal. It took me many years to get to this level of simplicity, and I think I've taken that as far as possible now. So, from here on, I'll move back into more florid counterpoint with a new perspective.

Finally, at twenty-five, we get the subject in the lowest octave of the orchestra, and all voices move stepwise except for the characteristic leaps in the subject and primary countersubject, with but one exception: The leap down from a-natural to e-natural in violin one. I did this to set up the voice's exit, and to get the spicy major second at the top of that voicing (E-natural versus d-natural).

So, the exposition proper is twenty-nine measures in length, and then we get a four-voice version of the previous episode starting at thirty. This gives violin one the opportunity to bow out very gracefully, as the a-natural seems to be absorbed into the e-natural of violin two (Thanks in part to the setup I mentioned perviously).

Since we have heard this episode twice now, it is necessary to add a new element, and that is the six-measure interlude that starts at thirty-four. Again, everything is smooth and stepwise except the bass and contrabass lines. Basically, all leaps are in the characteristic leaps in the subject/answer and countersubject/counter-answer, or with note values less than a half-note with the lone exception previously pointed out.

The half-cadence into measure thirty-nine is reached via an augmented sixth, and here is where I modified things for the v1.1 I mentioned: Previously, I had the cellos go down to d-natural on the final beat of thirty-nine - to get a full V(m7) sonority - but it actually works better with the pause on the half note before the development section starts. I'm still trying to shed the vestiges of the traditional view that tonal fugues should have a constant surface rhythm, which is why I have consciously worked back to the Palestrina style in that regard. Plasticity of line is the essence of that style, with a huge preponderance of stepwise motion.

Since the subject was composed as a five-part canon, there are myriad stretti possible. The simplest one that has a unique character to it is this one with two measures of overlap/three measures of delay. It is a perfect dovetail structure, as the countersubject does not have to be modified to lead smoothly into the subject - it is simply interrupted - and vice versa: The subject does not have to be modified to lead into the countersubject. My obsession with these kinds of dovetails goes back to my fascination with the art of M.C. Escher. I always wanted to make musical versions of these morphological structures, and so now I do. This whole fugue is based on different musical "Escher morphs."

After this simple stretto, I present a two-part version of the primary episode, which will set up a direct modulation to the dominant region for the three voice stretti. I hear this as a duet between oboe and clarinet, by the by.

Now we have a trio with flute, oboe and clarinet for the three voice stretti, but the clarinet is poised to bow out here after it presents a dominant pedal (From the perspective of the dominant region). The bassoon then enters in fifty-six, and for the first time we hear three measures of the subject simultaneously: Measures one, three, and five, from bottom to top. At fifty-seven the subject moves into the augmented version of the head, but this time for complete augmented statements. This is a true Escher morph, and the augmented note values mean this would be a stretto at a single measure of delay in the original note values. I've never encountered this before, so I may be the first to do this kind of mensural canon.

The three-voice stretto in augmentation then plays out to great effect, and that is followed by the head of the subject in double-augmentation starting in sixty-seven. The entire subject will not work in double-augmentation at two measures of delay - a half-measure of delay in the original note values - so I had to modify the tail. Measure seventy-two was the penultimate measure into the four voice stretti previously, but as I said at the end of the previous post, I wanted to let the doubly-augmented version play out in its entirety, and so that's what I ended up doing.

The doubly-augmented version slows things down to a very cerebral pace, and all I had to do was keep the rhythms, and ditch the chromaticism. This is exactly the same version of the subject that appears in augmentation in the final five-voice perpetual canon that ends the piece! Because of that, we get some very piquant minor ninths in the counterpoint: b-natural up to c-natural in measure seventy-three, and b-natural up to c-natural again in measure seventy-five. I feared this wouldn't work, but ended up loving it.

As in the conclusion, things re-dovetail back here, so this is a perpetual canon as well. Unfortunately, the subject in the original note values will not work, so I used the doubly-augemented version again and stopped at ninety-eight, on the exact same penultimate voicing as before.

At ninety-nine, then, the four-voice stretti begin, which will lead to the tension climax of the piece. The entries are still at two measures of delay/three measures of overlap, but I introduce a new element; the two-and-a-half measure tonic pedal that goes into a half-note of leading tone starting at one-hundred-two in the second violin part. This is then picked up in turn by the contrabass, the cellos, and then the violas.

The second violins then play two thematic statements in as row, as do the double basses, and so now the stretto is at a single measure of delay. Things are not in the proper order, however, so we end up with the basses and cellos landing on a unison f-natural in one-hundred-thirteen. The cellos then dip below the basses on the last triplet, and there is briefly a simultaneous f-natural and f-sharp. This is the tension climax of the piece. The cellos then break the canon with the augmented head figure starting in one-hundred-fourteen, and the tension recedes enough by one-hundred-sixteen to return for the final episode an interlude.

We then get the same episode and interlude that ended the expository section and introduced the development area, in exactly the same voicing: Setting up the development area so that it leads smoothly into this borders on the magical. I really don't know how to describe the thought processes that lead to something like this, as they are equal parts intuition and symbolic non-verbal thinking. In any event, it amazes even me. lol. The interlude is extended by a single measure so that it makes an authentic cadence versus the previous half-cadence. This sets up the recapitulation, which is introduced by the diminished triad triplet on the final beat of measure one-hundred-twenty-seven.

The recap is a five-voice perpetual canon. Measures one-hundred-twenty-eight through one-hundred-thirty-two are how I composed this subject almost exactly ten years ago. Then, the augmented form starts at one-hundred-thirty-three, and it is exactly the same as the doubly-augmented version from the three part Escher morph (This is just a five-voice version of the same morph starting with the original note values).

After the augmented version plays out, everything re-dovetails to the subject in the original note values, and the perpetual canon is proven.

After the canon plays out, the piece winds down over an ostinato on the final measure of the subject. The piece then ends by alighting momentarily on a major triad (Diad, actually), and then the final triplet has as the penultimate note, the minor second degree, which is the only pitch class missing from the subject.

Bada bing, bada boom.

I feel the need to drink a lot of Guinness. lol.

Saturday, March 09, 2013

Five-Voice Ricercare for Orchestra, v1.0

Well, what do you know. Less than a month after getting the five-voice exposition done (previous post), I got a first version of the development area finished.

This is huge, as this piece is probably obviously the finale of a symphony. It also just may be that this ends up being the final version, as it is well proportioned and filled with clever contrapuntal devices that are quite dramatic. I've been listening to it for several days, and for now I don't see any way to improve upon it (These things are always subject to change, of course).

Here's today's MIDI to MP3 conversion: Five-Voice Ricercare for Orchestra, v1.0

Once through the exposition again: On the top system is the five-measure subject, which I composed as a five-voice canon at the octave and single measures of delay. With chromatic subjects like this, the answers have to be real, and so the real answer and its counter-answer are on the second system. Starting in measure eight, the counter-answer is the head of the answer in augmentation.

The augmented head of the answer in the counter-answer makes the expected leap of a diminished seventh into the next statement of the subject. This works out perfectly because the minor sixth degree from the perspective of the dominant region - the C in the viola part in measure eleven - is the minor third degree from the perspective of the tonic.

This doesn't work out going from the tonic to the dominant, however, as the minor sixth degree from the perspective of the tonic is the minor second degree relative to the dominant. So, I had to insert the four measure episode you get at sixteen to reorient the voices for another dominant statement. This greatly increases the interest of the exposition.

The real answer then enters again in measure twenty, and in the highest octave of the orchestra. The C-natural in twenty-one is, in fact, the pitch climax of the entire piece. At first blush, it may seem strange to have the pitch climax in the exposition, but actually, if you want to take a Schenkerian view of it, the piece has a 5-4-3-2-1 line through it, as the real answer on the dominant degree becomes the subject on the tonic degree in the recapitulation.

Then the final answer appears in the lowest octave, and it hits the lowest note in the orchestra - the E-natural in twenty-nine - a mere eight measures after the peak, and five octaves plus a minor sixth lower. That was the point of this fugue: To write a five-voice fugue that used all five octaves of the orchestra (Or string choir, at this point).

A larger version of the previous four-measure episode then returns at thirty, and then a new element is required, which is the six-measure interlude at thirty-four. I changed the ending versus the version in the previous post, as it must now lead into the middle entries instead of the five-voice perpetual canon that is the recapitulation. This has a less than perfectly satisfying effect, which will be, "fixed" the next time we hear this.

Our middle entries begin with a stretto that is also a perfect dovetail structure: The counter-subject does not have to be modified to enter the subject - it is simply interrupted - and vice versa; the subject moves seamlessly into the counter-subject (The augmented head of the subject at forty-five). I then present a two-part version of the four-measure episode at forty-eight, and we're set up for the three voice section. This was taken directly from the two-part solo guitar version, and I hear this as an unaccompanied duet between oboe and clarinet, by the way (I will introduce rests when I get to orchestrating).

Whereas the previous stretto had two measures of overlap/three measures of delay in the tonic, these will have three measures of overlap/two measures of delay on the dominant. Here I present a sort of mensural canon I've never encountered before, and for which I have no appropriate terminology. The subject statements at two measures of delay dovetail into augmented versions of the entire subject at two measures of delay, which would be one measure of delay in the original note values. It is very, very cool.

As this unfolds, the tempo drops to almost nothing, and since the head of the subject will work at a half-measure of delay in the original note values, I can introduce it in double-augmentation here (Starting at sixty-seven in violin one). This is really an amazing passage! Years of working with contrapuntal dovetails has lead up to this; one of the coolest dovetails I've yet devised.

Well, the good old diminished seventh leap in the doubly-augmented version allows another driect modulation back to the tonic, where I present the four-voice stretti. This section is obviously highly organized, but it is weirdly dissonant and unsettled: Things work out fine technically, but they are not in the most copacetic possible order.

Starting in eighty-three we get the tightest possible stretto; that of a single measure of delay. This too has an unsettling effect, as things are not in the most perfect arrangement, which raises the question, why? This aggressively strident section peaks in measure eighty-seven, where the cello and bass land on a unison, after, "kissing" many times on seconds. The cello then dips below the bass on the final triplet of that measure. An awesome and startling effect, but why?

As the augmented versions begin to appear in eighty-eight, the tension reduces, and all is calm again by ninety.

All questions are answered here, as we reenter the episode and interlude that were at the end of the exposition, with nothing needing adjustment and all voicings the same. To set up a highly organized musical drama of this intricacy is obviously a very deep thing. There's so much that can't be put into words, as the thinking that leads to this is symbolic, and not verbal. One thing is for sure, and that is that it's awesome.

In one-hundred and one-hundred-one I, "fix" the interlude so that it resolves properly, and that leads into the five-voice perpetual canon that is the recapitulation and codetta.

Here, at last, everything is presented in the perfect order, and everything leading up to it makes perfect sense now. Violin One gets the subject on the proper level instead of the answer a fifth above, and so the thematic resolution is complete.

I've described this perpetual canon many times before, so I'll just let it speak for itself now.

And so, there it is. Epic.

The thing I'm considering right now is to allow the doubly-augmented version play out, which is possible with the diatonic augmented version here in the recap. On with the show, then.