Friday, August 12, 2011

Après-MIDI

I've been mulling over topics for new posts, many based on conversations with other composers, students and performers about how composers are taught, how we learn, the many ways in which composers produce music. But one topic jumped out at me in the last day, and I bring it up here as a way of pondering its significance for my compositional process, and perhaps that of others.

A Finnish composer friend who now lives Stateside came over for dinner last night. He's a very articulate, very opinionated guy, just my kind of discussion partner. As usual, we immediately got into an evening-long debate about aesthetic values, composition training, repeating oneself artistically, and other topics we always seem to gravitate toward. Comparing our recent projects, I mentioned that I was currently about two-thirds of the way through a new piece for the Zagros ensemble, and it was the first time I'd composed a piece entirely with MIDI. It's an experiment I'd been conducting to see how I fared using this newfangled tool. I've admitted in the past, and proudly so, to being a very old-fashioned composer. Until now, I've hardly ever touched the playback feature on Finale/Sibelius while in the act of composition. I usually write everything by hand, from sketch to full score, using the computer for engraving only, much to the amazement of a lot of my colleagues, and especially my students. It's not so much a matter of habit, as that I feel I actually work faster this way, especially when I'm orchestrating. Being able to see the entire page, or twenty of them in a row if I want, being able to move my hand around freely and just scribble notes as the mood strikes me, gives me more freedom. While I enjoy working this way for the most part, I do admit to a certain amount of envy at the technical fluidity of my younger colleagues, who all seem to compose directly into the computer these days. To them, I imagine I seem quaint dragging around my huge architect's portfolio with my hand-written scores in them. (If you want to know real fear, carry your only manuscript tabloid-size copy of a 35-minute orchestral score around a city on public transportation.)

There are other reasons for my stubbornness and reluctance when it comes to integrating the computer into my composing routine. I've always looked at computer-based composition of acoustic music with a jaundiced eye. First of all, there was the sound quality of the playback. When Finale first came along, the sound was so bloody awful that I couldnt' stand it. So bad, in fact, that I couldn't even bring myself to use it to check the pacing of a section with it, because the tinny, awful approximations of acoustic instruments just ruined my sense of the harmony (back when I used such a thing), articulation, phrasing and dynamics. It was like trying to read with a strobe light in your face. Now, of course, the built-in sounds notation software packages come with have vastly improved, with some of them, woodwinds especially, being quite convincingly lovely. If they could just kill the James Galway-esque vibrato on the samples, I'd much appreciate it. Although I've recently divorced Finale, I admit that its human playback feature is pretty helpful in some circumstances. But still, as I grew to incorporate a modest number of extended instrumental and vocal techniques, and to use freer types of notation in realizing a texture, the software didn't. I was therefore left with no choice but to just feel my way by ear toward the result I wanted.

Then there's the impact of notation software on the compositional process itself, and the playability of the result. In my experience, one can usually tell when a composer uses the software to compose directly. There tends to be a marked favoring of cut-and-pasting of entire sections of music, entire patterns, wholesale transposition of harmony in favor of voice-leading. Not that any of these devices is bad in and of itself, but the medium tends to encourage a kind of compositional laziness in terms of creation and manipulation of material.

A side-effect of using MIDI playback in composing, I've noticed, is a lack of real-life playability of a lot of the rhythms. I've even noticed it in my own use of it. You can make a computer play back whatever you write accurately, but this can lead to the creation of irrational rhythmic layout whose goal is accurate playback by an electronic processor, not by a human. Of course, any decent composer will know about and correct this, but many don't. In this last piece, I've gone through it with a fine-toothed comb looking for things like downbeats that I wrote as offbeats to get the pacing right on MIDI, fast rhythms that might be better interpreted as grace notes, strict, complex rhythms in melodic writing that would be easier to read, and sound better if I were to relax and simplify them.

In the field of playability, instrumental or vocal idiom is also an issue. However good the human playback protocols of whatever program you're using, they just can't replicate the subtle difficulties of the physical technique involved in playing an instrument. Although such features can be useful in reminding one just how long it really takes a contrabassoon to sound its lowest notes after the initial attack, they can't tell you about how it feels to play through a difficult passage. The computer will happily play back whatever you tell it to, irrespective of whether or not it's really possible, which can lead to radically overestimating and instrumentalist's or singer's ability to realize what we write for them, disregarding their comfort in the realization of our compositional vision. Again, any composer worth their salt will compensate for this. Far too many don't.

Then there's the more intangible category of MIDI's effect on the inner ear, the composer's sense of instrumental or vocal sound. Frequently, when listening to a new piece, I can tell if it was orchestrated on MIDI. First of all, instrumental balance in MIDI is not, and probably never will be anywhere close to reality. When a piece of chamber or orchestral music has balance problems, it can usually, in my view, be attributed to its having been scored on MIDI, where any problem of volume or attack can be overcome by giving it more cowbell on the mixer. Learning orchestral balance takes a lot of time and experience. Not having had a terrible load of experience writing for orchestra early on, I attribute my success (so far) in the medium to my previous life as an orchestral/ensemble performer. Those years sitting at the back of the band with a score were the best education a young composer can get as to the inner workings of the orchestral beast, how a section of strings sounds at various dynamics, what kind of articulations produce the best results in the brass, etc. Sitting in choirs since I gave up instrumental performance has been equally beneficial to my choral writing. Spending years working with an ensemble from the inside, as a performer, has given me a bone-deep sense of what works and what doesn't. None of that is possible working on a computer with synthesized or sampled approximations.

Subtler still is the quality of orchestration MIDI yields. Again, when a piece has been orchestrated on MIDI, I can usually tell, because there's a marked lack of invention in the orchestration. (This is not to accuse anyone who works this way of being unoriginal, it's more of an indicator of my inability to come up with a better word.) MIDI, however much it's developed over the last decade, is still a very traditional, hidebound orchestral tool, meant largely for writing commercial music with simple orchestral solutions. Working directly into MIDI as an orchestrational tool encourages, in my view, obvious solutions in distributing orchestral material. The resulting music sounds like MIDI. It behaves like MIDI. The interaction of instruments is less idiomatic and timbre-based, and more about how easy it is to move material around in an electronic setting. There's very little sense, frequently, of the composer reaching for a sound, and as a corollary, there's less risk of failure, and failure at realizing an idea is a major component in creating a new or fresh sound. This is, in my opinion, the worst aspect of working directly with MIDI, the way it seems to dampen the imagination in terms of sound creation. What MIDI can't ever replace is a sense of "what if?". What if I did it this way? Wow, I've never heard muted piccolo trumpets and piccolo together, I wonder if that would work? How would the texture sound if I had a harp behind that string tremolo? These things are replicable to a certain extent in the computer environment, but will never sound real enough to give an accurate impression, so we're left to simply imagine it, to write it down and hope it works. The risk of utter failure has to be part of orchestration. As one of my previous teachers said, if you write an orchestra piece and it doesn't turn out exactly the way you expected and you're surprised by that, you're a fool. And if it does turn out the way you expected, you're a lucky fool. I'll take luck over certainty any day. "Huh, that didn't work out the way I planned, but it's still pretty cool."

Which brings me to my own recent inclusion of MIDI in my process. I started using it a lot to make quick mock-ups of sections of my viola concerto to check pacing of rhythm and pitch-field turnover, as well as to check the heterophonic counterpoint of a large section of polytonal melodic writing. The reasoning behind this move was that it was a huge piece, and a huge, very public opportunity, and I didn't want to mess it up when I could have checked these basic things in a controlled environment. I'm not too proud to admit that MIDI saved my ass at several points in the piece when I just couldn't work out things on paper, or had radically underestimated the pacing of a series of phrases. But I never – and mark this – ever turned to the computer until I'd figured out the entire pitch content and essential texture of a section. In a sense, the music wasn't composed at the computer, only arranged there. Wondering why I resisted integrating the computer into my routine so much, I realized that it was because I usually do a lot of the initial pitch work at the keyboard, or on whatever instrument I'm writing for if I can get my hands on one, and if I didn't have an instrumental interface, I lacked the confidence to just throw notes around on the screen. I'm not the sort of ultra-musical prodigy type who has music pouring out of every orifice and seems to just conjure things out of thin air with blindingly fluent technique. I'm man enough to admit the very real limits of my talents, and my very real attachment to sound in music rather than its technical or linguistic features, and make allowances in consequence. Getting the notes right, even if it's just five pitches, is a huge, time-consuming, doubt-ridden, long-dark-night-of-the-soul part of the process for me.

So as an experiment, I decided to buy a small MIDI keyboard and write this current chamber piece directly into the computer, just to see if I could adapt. So far, I'm finding myself comfortable doing it. It's certainly very handy to have quick access to all the instrumental sounds through the keyboard. It's an odd ensemble I was asked to write for: flute, clarinet, bassoon, harp and string quintet. There are tricky balance issues in play, made more so because they asked me to somehow spatialize the material so the wind players could move around the art gallery where the premiere will take place. But as a group of single instruments, the balances sound more realistic on MIDI, so it seemed like the right forum to conduct this trial in.

The resulting music is quite simple, very much like Feldman's Rothko Chapel in the way the materials relate. Not wanting to create a kind of mini-orchestra, or the obvious mini-concerto for harp (the ensemble is basically an extended version of the one in Ravel's Introduction and Allegro, which I know well from hearing my wife learn it last year), the instruments never play together at the same time. It's more a non-linear sequence of solos and ensembles, quite simple for the sake of coordination across a big space, but also because I'm deliberately keeping it simple for myself as a compositional exercise. I did, however, do some things I would never have had the guts to do on paper, like a section of Ivesian multi-stylistic counterpoint, where the wind trio babbles in rather banal atonal counterpoint over a diatonic bed of strings. I wouldn't have tried it without the playback as a guide, because while I could guess that the general effect would work, I couldn't be sure it would sound right. As I don't have a particularly good ear for contrapuntal structures, MIDI gave me more confidence in working out the idea, a greater sense of certainty, and allowed me to get the notes right where I couldn't have done it by ear on paper.

And that sense of certainty is what brings me back to my conversation with my friend. I told him that usually at this stage of a piece, my insomnia has gone into overdrive, with ideas looping in my head on infinite repeat as my mind works on the material, where to go next. But this piece isn't keeping me up at night. I attributed this in our conversation to my just being tired and a little burned out creatively. A part of it, to me, was the non-linear nature of the ideas. I'm not thinking about causal connection or development or transition because I don't want there to be any. I just sit down in front of the computer every day and think about where I want to go next, on a very basic, very intuitive level, like putting together an art exhibit rather than a single piece. My friend was kinder, offering that maybe it was because I've gained sufficient control over my process and critical detachment from the work that I'm able to walk away from it more easily, confident that I'll find my way back in if I don't think about it all the time. This has a certain validity to it.

But as I was brushing my teeth after yet another uncharacteristic night of sound sleep, the thought suddenly hit me: what if I'm not pondering this piece because in the very act of using MIDI playback as a tool, the music no longer holds any sense of mystery for me? I know the pacing of ideas is right because it sounds right, here and now. I don't wonder how two layers of contrasting material will sound because I know how it sounds, and it works. I don't need to worry about how my disconnected, stream-of-consciousness series of little episodes works because I know (or at least think) it does. What if the worst thing MIDI does to us as composers is kill our sense of fancy, our need to worry over the result? Why worry about a piece if you already know it's going to be successful? Working in real time, with a reasonably approximate result of our work available to hear at the click of a button, do we lose a sense of possibility? Do we lose our ability – our willingness – to fail?

I should follow this Sex and the City-style voiceover with the caveat that I don't think using MIDI as a major compositional tool is in itself wrong, or that it yields bad results, or that it inherently makes composers lazy. If nothing else, it makes bad or already lazy composers think composing is easier than it is. (This applies to other software as well. Open Music, it strikes me, is equally dangerous, albeit more sophisticated in its cut-and-paste, idiom-negligent potential.) Any good, sensitive, musical composer will be aware of these problems and correct them. In the end, there is no right amount of computer use in composing. Everyone has a different ratio that suits their needs. I've found a few limited ways of making it useful to me. I don't think I'll ever transition to full computer use in composing, especially in orchestral music, because I know that my ear and experience with the orchestra are pretty reliable, and generally more fine-tuned. But in the integration of computer playback into a compositional routine, do we become reactive rather than proactive in shaping the outcome of a piece, of a sound? There's a slippery slope here, one which I find myself sliding down despite my best efforts to remain aloof. Yet the lure of total certainty is a powerful narcotic. I wouldn't be the first ascetic to succumb to Bacchic debauchery in the name of maintaining the appearance of infallibility. And that's what the canonization of "great" composers and the music marketing world teaches us, that we need to appear infallible, assured, demigod-like in our creative powers? Where is the point of no return? I suppose that's for another day. For now, I still need four minutes of music on this thing.

Click.