American Composers Orchestra


about the concert


aco homepage

concert schedule

Related Essays:

The Copland-Sessions Concerts: A History
by Carol J. Oja

Copland & Sessions: A Musical Friendship
by Vivian Perlis

Modernizing Ballet Mécanique:

'90s technology makes a '20s piece possible

By Paul D. Lehrman

There was a lot about the original1924 version of George Antheil's Ballet Mécanique that was shocking, and high on the list was the fact that it called for, along with two grand pianos, three xylophones, four bass drums, and a tam-tam,16 synchronized player pianos. While the player piano had been around for some 30 years, and other composers like Stravinsky, Ravel, and Hindemith had written or arranged pieces for the popular mechanical instruments, Antheil's youthful magnum opus was the first concert work to call for multiple player pianos. Not to mention the sirens, bells, and airplane propellors.

Antheil knew he was breaking new ground, but he didn't know how far into the future he was reaching. As far as he knew, the technology to play the piece existed. Pleyel, the Paris piano maker whom Antheil worked closely with (and whose concert hall, Salle Pleyel, was to be the site of one of the first public performances of the work), had designed and even patented a system for synchronizing multiple player pianos (or "pianolas"), using electric motors and a sophisticated speed-adjusting system that bears a striking resemblance to the timecode technology used today to synchronize electronic sound and images.

Unfortunately, it appears that Pleyel never actually built a working system. Antheil and his friends made up all sorts of excuses as to why the premiere of the piece was being delayed (Pleyel was having trouble with the rolls, there wasn't enough electricity in any hall in Paris to power that many pianolas, etc.), but the truth was that there were simply no instruments in existence that could play the piece. When he finally realized this was the case, Antheil re-wrote the piece for one pianola and multiple human-played pianos, and that version was heard several times in Paris and then in 1927, in a concert noteworthy for its utter failure, at Carnegie Hall.

Today, thanks to computers, the Musical Instrument Digital Interface (MIDI), and MIDI-compatible player pianos like the Yamaha Disklavier, Antheil's youthful vision can finally be realized. MIDI is a computer language that allows one electronic music device to control another-in its simplest form, it lets a musician play on one keyboard and have the sound emerge from another instrument. MIDI is ideal for synchronizing large numbers of musical instruments with great precision-it is used as a composing tool by almost all musicians in the pop field, as well as in Broadway shows and Hollywood film scores, and even by many present-day composers of concert music.

Antheil's score called for four pianola parts, each to be played on up to four instruments. My task, which I took on at the behest of G. Schirmer, was to take these parts and transcribe them into a MIDI "sequence" file. This is a computer file that contains all of the note, timing, and other musical information in a score, which can be used to play the score on any MIDI-compatible instruments. I entered the music from the printed score prepared by Schirmer into a Macintosh software program called Vision, a MIDI sequencer made by Opcode Systems. MIDI sequencers can cut and paste musical data much as a word processor can edit words and paragraphs, and this was a great help in dealing with the repetitive parts of the score. In addition, sequencers allow you to enter note data at any tempo at all, and then change the tempo afterwards. Given the density of the pianola parts, this was also an important feature.Yamaha Disklaviers are marvelous instruments, but they were not necessarily made to play the type of music Antheil envisioned.

Disklaviers are designed to accurately reproduce the playing technique of humans, who have ten fingers. The instruments can play up to 16 notes at a time, which normally would be considered generous. Antheil's score, however, calls for chords of up to 23 notes. Fortunately, in those places where the chords are that thick,they are doubled in all four of the pianola parts. Therefore, it was possible to thin out the chords, removing different notes from each part, to bring them within the Disklaviers' capabilities, but without sacrificing the sonic impact of the passages.

Antheil also called for very fast repeating notes and chords, something humans don't normally do, and therefore something Disklaviers are not particularly good at. After a substantial amount of experimentation, I was able to find ways to get the Disklaviers to respond to these passages. Sometimes it involved judiciously thinning chords, while other times it involved changing note durations and key velocities in such a way as to essentially trick the Disklaviers into doing things they normally would refuse to.

One other interesting aspect of the Disklaviers is that when they receive incoming MIDI data from another keyboard or a computer sequencer, there is a delay of exactly 500 milliseconds before they respond. This is not a fluke: it is actually a very smart decision on the part of the instruments' designers, and it is necessary if multiple recorded keystrokes at different key velocities are to be able to sound simultaneously. But it means that if the sequencing program playing the Disklaviers is also controlling anything else (and it is in this case, as you'll see in a moment), the Disklavier "tracks" have to be advanced-literally moved earlier in time-by 500 milliseonds.

Besides the pianolas, Antheil's score calls for three airplane propellors, seven electric bells, and a siren. In order to make the piece as easy as possible for musical groups to perform, I assembled digital samples of all of these sounds, and they are provided along with the score on a CD-ROM. The airplane samples were recorded by Tim Tully at a private airfield in northern California; the siren was recorded by me at the Arlington, Massachusetts, fire station, and the bells were recorded in my studio. The samples are supplied in several formats so that they can be used with a variety of commercially-available digital samplers. Samplers are electronic instruments that play back pre-recorded (as opposed to synthesized) sounds in response to incoming MIDI data, either from a keyboard or from a sequencer. The sequences containing all of the pianola parts that I created for Schirmer also contain tracks for all of the sound-effects parts. Performing groups have a choice: they can use the sequenced tracks and have the effects play automatically, or they can use live MIDI keyboard players to play the samples.

Yet another option is available when it comes to the bells: in preparation for the first performance of this piece, my colleague Coleman Rogers and I designed a contraption comprising seven different-sized electric bells and a series of relays, controlled by a MIDI-to-relay convertor made by MIDI Solutions. This "MIDI Bell Box" is designed to be hung above the stage, and it too can be played automatically from the computer sequencer, or by a live musician playing a MIDI keyboard.

Finally, the MIDI sequence also contains a "click track": an audio track containing percussive sounds similar to a metronome. This track is fed to the conductor through an earpiece, and serves to keep him or her on track through the wildly varying time signature changes (there are over 600), as well as the long silences. The track also provides audible cues of rehearsal letters, and even "countdowns" to important transitions. Generally, the computer will determine the tempo, and the conductor will follow, but it is certainly possible to perform the piece with a human adjusting the tempo of the sequencer in real time, following the conductor. Perhaps even more intriguing, it is possible that all of the performers could be wearing earpieces and listening to the click track at the same time-and this could obviate the need for a conductor entirely! Somehow, I think George Antheil would have gotten a kick out of that.

Paul D. Lehrman is a composer, author, and consultant. He is on the faculty of Tufts University, and is editorial director of the World Wide Web site for Mix magazine, the leading journal of professional audio. More information about the Ballet Mécanique can be found on his Web site

aco homepage

concert schedule