I've had a couple of opportunities to play my Disklavier pieces lately, in New York and at Bard. A Disklavier, just to be very clear since so many get the wrong idea, is an acoustic piano, with real strings struck by felt hammers and vibrating in the air, but operated from a computer (or disc) via MIDI instructions. The keys move, just as though a pianist were playing them. It's a modern player piano, only the paper piano roll is now replaced by a sequence of digital information.

Anyway, the response I get is kind of deadeningly repetitive. The pieces I usually play are jazzy, impressively fast, and sort of humorous, and generally make a good impression. But afterward, I'm invariably approached by two or three or four people who ask, "Gee, isn't there some way to make it possible for live pianists to play those pieces?" They ask as though they suspect it's a possibility I've never considered, as if they expect me to strike my forehead and shout, "Of course - a live pianist! Why didn't I think of it?"

Now, number one: I get a big kick out of watching the Disklavier. It's fun to watch all those keys ripple up and down the keyboard; I take the front cover off, when it's an upright, and you can watch the hammers y by as well. In Australia I also hooked up my computer to a projector, so the audience could watch the Digital Performer le scroll by, which looks exactly like a player piano roll, only with the notes running horizontally instead of vertically. I got the idea from watching Conlon Nancarrow's player pianos, which were incredibly more fun to watch live than to listen to a recording of. You'd see a diagonal line of holes appear on the piano roll, and know that a huge glissando was coming, and it would blast in a split second later - it was like being on a sonic roller coaster, because you could see what you were headed for just a second before it happened. I love watching player pianos as much as I've ever loved watching a live pianist.

Number two: I've written a lot of piano music and a lot of Disklavier music, and I approach them with different mindsets, just as though they were different instruments. When writing for Disklavier I don't even think about spacing the notes so that a human hand can reach them. If I want to write a melody in lightning-fast quintuple octaves, or a whole string of six parallel sixths, I go right ahead. And the whole point is to be freed from downbeats and meters, so the rst thing I'll do is lay out a whole set of nested tempo relationships, like 7-against-9-against-11-against-13-against-17, and then ll in the notes, knowing that notes in one line will coincide with notes in another line only at downbeats, and then I try to avoid putting notes on downbeats. By doing that I get exactly what I want, which I feel is a wonderful spontaneity of notes bubbling up, not randomly, but like corks bobbing up and down on brisk waves, with patterns that are repetitive but wholly unsynchronized.

I know that there are pianists, like Ursula Oppens, who have trained themselves to play some pretty complex rhythms; in fact, the Helena Bugallo-Amy Williams Piano Duo played some of Conlon Nancarrow's early Player Piano Studies in New York this past Thursday, and I couldn't be there because I was impersonating Abraham Lincoln that night. (Scroll down if you really have to know why.) But it's one thing to play a 22-tuplet over a 4/4 beat in a Chopin nocturne, it's something else to play steady lines of 13-against-29-against-31 for several measures at a time. I imagine it can be done. What I don't imagine is that it would sound the way I want it to sound, with the same spontaneity and bubbly effect. I did, by request, transform one of my Disklavier pieces (Folk Dance for Henry Cowell) into a live-pianist piece (Private Dance No. 2), and I've never been totally convinced by the result. In addition, pianists fudge rhythms like these, and I frequently change harmony in mid-measure among several lines at once, the notes all changing chord suddenly like a flock of birds mysteriously reversing course with one mind. I don't see how a pair or trio of pianists would be able to "sort of" play all these tempos at once, and also be able to so closely synchronize that when one switches to the E minor triad on the fifth note of a 17-tuplet, the other switches to that chord on the corresponding fourth note of a 13-tuplet.

Maybe it could be done. If someone can figure out how to do it, I'll applaud. But the other thing I can't understand is, why would anyone want to go to that much trouble? Why are so many people so dissatisfied watching the Disklavier, even people who visibly enjoy it? Sometimes the question comes from a pianist who is dazzled by the music and wants to play it, and that's flattering. I wish I could interest these pianists in the eight or so piano pieces I've written for human players, but I rarely do. One person said that the Disklavier doesn't give the feel that a live pianist can. Well, that's a point, I guess; but unlike the old player pianos, I can adjust the dynamic (hammer velocity) separately for every note, and I do a tremendous amount of ne-tuning to accent just the right note in a phrase, humanize the attack points, create the effect of a live pianist hesitating on a high note or beginning a trill slowly. I simulate live performance with what strikes me as a high degree of realism, and I am strongly tempted to assume that psychology plays a role in perception here - the music often sounds nuanced, tentative, slightly irregular just the way a pianist would play it, but since there is no pianist, the listener fools himself into believing that it sounds regular and mechanical.

Or is it just that people don't enjoy watching machines play music? I've seen an entire orchestra of MIDI-operated machines play music in Trimpin's studio in Seattle, and it was one of the great musical thrills of my life. Computer-operated acoustic instruments are coming, folks - they're part of your future. Get used to 'em or ignore 'em, but you can't stop 'em.

I know Conlon used to be bothered by similar queries. In his day, there was always the complaint (he got it from Aaron Copland, among others) that a player piano performance was the same every time, that there was no interpretive deviation from one playing to the next. Conlon's usual response was, a Picasso painting is the same every time you see it; a Shakespeare sonnet is the same every time you read it; why is only music required to be different every time or you can't enjoy it? Today's audiences, however, are so inured to recordings and even near-identical performances that that objection seems to have disappeared. But for some reason people are just bothered by using a computer to do something that humans have always done, and they seem willing - as I am not - to put up with any compromise to transfer that activity back into the traditional realm of the performer-audience relationship. I wish I understood why. Because, sadly, I think people who strongly feel that way are just going to have to listen to someone else's music, and there's a lot out there.

Copyright 2004 by Kyle Gann

Return to the Kyle Gann Home Page

return to the home page