Disklavier FAQs

In response to my new CD Nude Rolling Down an Escalator the questions have started pouring in about the Disklavier, some of them the same questions that Conlon Nancarrow spent his late life fielding about the player piano. Let me see if I can head some of them off at the pass.

I love the pieces, too bad the Disklavier sounds so electronic. Couldn’t you have used some really good piano samples? Actually, the Disklavier is a regular acoustic piano. Those are physical, metal piano strings being struck by felt hammers, just like any other piano. I can reach in and pluck the strings if I want. It’s exactly like an old-fashioned player piano, simply played by MIDI commands rather than by a paper roll with holes in it. If you think it sounds electronic, your false conception of what a Disklavier is may be misleading your perception.

The one odd thing about my Disklavier is its tuning: I keep it in an 18th-century well temperament, Thomas Young’s well temperament of 1799 (nearly identical to what’s called Velotti-Young on some synthesizers – you can read about the scale here). It’s a more subtly different tuning, to our ears, than something like Werckmeister III that Bach used; the greatest deviation from modern equal temperament is only 6 cents (6/100ths of a half-step). It is not a “microtonal” tuning, as some have thought, because there are only 12 pitches to the octave, all about a half-step apart. Nevertheless, while it’s difficult to notice the well temperament in any particular passage (though one reviewer’s sharp ears caught it in Folk Dance for Henry Cowell and Tango da Chiesa), it does create a slight but pervasive difference of timbre over the whole keyboard. Intervals that are purer, and lack the buzzy inharmonicity of the modern piano, are often perceived as unpianolike, and a little bell-like or electronic. I’ve had this perception myself with La Monte Young’s The Well-Tuned Piano. If you think the piano sounds a little electronic, it might be that you’re not used to the temperament. I recently had my grand piano at home worked on, and it came back in equal temperament; I couldn’t stand the sound, which was buzzy and harsh and undifferentiated, and which everyone else is perfectly accustomed to. I was so relieved when my piano tuner came over and restored the 18th-century temperament.

[UPDATE: Composer Lawrence Dillon credits the electronic illusion to "an aural illusion caused by fast torrents of notes that I intuitively knew couldn't be contained in 10 fingers -- my brain... solved the riddle by hearing an artificial tint to the timbre." I have to admit, there's a moment at the end of Bud Ran Back Out that sounds electronic even to me. Perhaps instead of worrying about that I should cultivate it.]

The tempos sound so mechanical – shouldn’t you have randomized the attacks to make it sound more like a human is playing it? Actually, I randomized the attacks in every piece. Almost nothing on the CD is metronomically pure. Again, if you know you’re listening to a machine, you may be predisposed to hear it as mechanical. However, in writing music of different tempos, there’s a limit to how much rubato one can allow, and it is a much narrower range than is common in live performance. (This is a question that came up constantly regarding Nancarrow’s player-piano music, and my feeling about it is the same as his.)

Jonathan Kramer, in his book The Time of Music, reported that studies that analyzed performers playing conventional music showed that even the most accurate performer will frequently show variation in the durations of consecutive 8th-notes or quarter-notes of as much as 15 percent. One study showed that professional violinists played a 3/4 rhythm of alternating half- and quarter-notes at a ratio averaging 1.75:1. Now, for the kinds of tempo contrasts I use, and that Nancarrow used for the latter half of his output (up to 60:61), a 15-percent tempo deviation would be fatal to the subtle differences between lines. Take one of the simplest examples, my Texarkana. The tempo contrast throughout is 29 in the treble line against 13 in the bass line. The joke of the piece is that the melody, being indefinably just more than twice as fast as the bass, sounds out of control. 26 against 13 would be a pedestrian 2:1, something any human pianist could do. Yet a 26-tempo is only an 11-percent deviation from the 29-tempo, well within the range of typical human tempo deviation. For the 29:13 tempo contrast to mean anything, the random attack humanization needs to be kept well under 10 percent.

What Conlon always said was that, in Romantic music, performers had to add rubato and tempo deviations to enliven the music because it was inherently rhythmically uninteresting. In his own music, he felt, the rhythmic interest inhered in the subtle complexity of close-but-not-identical simultaneous tempos, and therefore no further “enlivening” was needed – and, in fact, would obviate perception of the tempo relationships he was trying to capture. I agree. To gain the new rhythmic liveliness of simultaneous tempos, we have to sacrifice some of the old rhythmic liveliness of rubato. Imagine if player pianos had always been around, but people had only recently learned to play piano by hand: someone would be complaining that we lost the old rhythmic liveliness of multitempo, for which pianists were fractically trying to compensate by applying rubato.

Performers have begun arranging Nancarrow’s player-piano studies for live ensembles. Don’t you really hope someone will do that for your pieces someday? Number one, just about the only Nancarrow studies that have been performed live are those with fairly simple tempo ratios, like 3:4:5. No one has yet arranged (or at least performed) Study No. 33 with its ratio of 2 against the square root of 2, or No. 40 with its ratio of e-against-pi. Similarly, I doubt that an ensemble could play the 29-against-13 of Texarkana, or the 5:7:9:11:13:15:17 of Unquiet Night. If someone wants to try, that’s fine with me – but it sure seems like a lot of wasted effort. Personally, I find both the player piano and the Disklavier tremendous fun to watch, whereas I don’t really see much entertainment in watching most live pianists.

The thing is, if you presuppose that the raison d’etre of a Disklavier is that it can do anything a pianist can do and more, I guarantee you’ll be disappointed. I’ll go further than that: if you expect ANY new music to provide all the same pleasures as the music you already love, I promise YOU WILL BE DISAPPOINTED. The question with new music is always, Does it provide sufficiently plentiful and rich new pleasures to compensate for the old pleasures that have been lost? A human pianist is an amazing phenomenon, and the Disklavier is no substitute for one; nor is a living pianist a substitute for a Disklavier. Each can do things the other one can’t. The fact that the sounds are the same may create an unfortunate expectation, one that’s never bothered me, but it may bother you. In some of my Disklavier pieces (especially Texarkana and Despotic Waltz) I take great fun in mimicking the conventions of live piano playing with the Disklavier, and, to me, it’s funny because they’re so not the same. I’ve written a lot of piano music for live performers, and I compose very differently for pianist than I do for Disklavier. To me, they’re different instruments. You may be one of those people for whom the Disklavier can only remind you of a deficient live pianist. If so, there are a couple thousand recordings of live pianists I can recommend.

Some of us composers feel that in order for music to progress, we need access to rhythms and tunings and timbres and structures that humans can’t play. Something will be gained by achieving them, but something else will be lost. I guarantee it. You’re either interested in the search for new musical pleasures or you’re not.

Why don’t you refer to it as the Yamaha Disklavier, since it’s made by Yamaha? Because I tried to get Yamaha interested in putting some money or publicity into the project and they turned me down. Why should I supply them with any more free publicity than I have to?

I needn’t have called them Disklavier Studies, after all, because they can also be played on a Pianodisc system. The Pianodisc system can be installed on a regular grand piano (a Steinway or Bösendorfer, for instance), and runs just like a Disklavier – with the additional advantage that Pianodisc, unlike Disklavier, can be run from a floppy disc containing straight MIDI files. Yamaha’s Disklavier ain’t the only game in town.

Tweet about this on TwitterShare on FacebookShare on RedditEmail this to someone