Player Pianos: Pianolas & Reproducing Pianos

The subject of Player Pianos, originally triggered by the discussion of Conlon Nancarrow’s music in the previous post, has led me to be in contact with Rex Lawson at the Pianola Institute in the UK, an organisation dedicated to the preservation and performance of these fascinating machines.

It turns out Rex was a good friend on Nancarrow’s in the 1980’s and 90’s, and Nancarrow was in the midst of writing a concerto for him to perform, a task unfortunately interrupted by his death in 1997. Rex was kind enough to provide a lot of information and correct me on my descriptions of the capabilities of the various types of Player Piano.

What I call the pianola, the foot-operated player piano, generally used rolls that were not recorded by anyone, at least not in the sense we use the word “recorded” nowadays.… Normal pianola rolls were transcribed from the score by musical editors, at so many fractions of an inch per beat, and any reasonable pianolist will play them differently from anyone else.…

The pianos which use recorded rolls are generally known as “reproducing pianos”, and in those cases pianists sat down at recording pianos, and some means was used for transcribing their keystrokes on to roll. Rolls for the reproducing piano play at a preset speed, whereas those for the pianola were intended to be varied at every moment, otherwise the music sounds mechanical, which was never the intention.…

Regarding Conlon Nancarrow:

Conlon came along at a time when the player piano was already part of history, and he used secondhand Ampicos, which were designed for the subtle nuances of pianists like Rachmaninov, as a quick way of bringing his own music to life. I love his music, and I play it quite a lot in concerts, but his compositions are not at all subtle from the dynamic point of view, because his main concern was to contrast different contrapuntal voices, not to phrase melodic lines in a romantic way. They are certainly not at all difficult to play on a foot-pedalled pianola. One only needs to be able to create a few dynamic levels, and the rest looks after itself.

And, on the subject of the Nancarrow Concerto:

Conlon wrote most of a concerto for me to play as a result of our meeting. He never finished it, but a British composer, Paul Usher, took all the sketches and made them into “Nancarrow Concerto”, which I did with a group called Ensemble Modern in Germany a few years back. There are plans to perform it in the UK in a couple of years.

More information can be found on the website for the Pianola Institute.

thinking aloud: listening and making generative music

Why just listen to generative music when you can easily make your own…1

It strikes me that because Generative music is likely different every time it is listened to, the act of listening to it can be accorded the role of creating that particular piece every time, therefore putting listening on a par with making, or even demoting listening as a role for the audience – maybe listening becomes “active,” listening as part of the generative process.

Generative music is the the most adaptable of “live” music – in that it’s always new, always different, no matter if it is a public or private situation, and it’s dependence on mechanical reproduction allows it the portability that “traditional” live music lacks. In terms of community experience, that particular version of the piece could be a shared experienced, in a live performance with an audience, but on a one to one basis it never has the potential to be communal, or shared – every launch of the piece occasions a new version, and unless it is hard coded as a static file cannot be transferred. But hard coding as a static file destroys the nature of the piece as generative.

One might ask what meaning it has if it is always different – where is the piece? This seems an inappropriate clinging onto outmoded forms, given the nature of generative systems, a paradigm for music (and indeed art in general) that may be difficult to maintain. If one must reduce it to such a thing one can look at it from two sides (maybe more). The meaning, the piece, is the method by which the piece is generated, the system, the order, the algorithm. Or, as I suggested above, the meaning is in the reception, but in a very fugitive sense. Static recordings can only be examples, maybe representing practical methods of revenue generation or dissemination, as in the production of documentation for ephemeral works by artists.

UPDATE: Coincidentally, Robin Peckham just posted a link to a piece by Nick Seaver, at the Comparative Media Studies department of MIT, where “he is studying indeterminacy and control in sound transmission, the role of ‘skill’ in aesthetic judgments, and the history of automatic musical instruments.” Nick posted a piece yesterday on his blog, noise for airports, about his research into the Player Piano. He makes an interesting point about “live” music, and its interpretation where the player piano is involved. This definitely relates to my notes above trying to define the “piece” where generative music is concerned:

“Live” music as we think of it today didn’t exist before audio recording—it was impossible for a sound to not be live. The player piano makes things a bit more complicated: is it still live if the notes are all punched on a roll in advance but “interpreted” by a live pianolist? Advertisements showing the ghostly hands of famous pianists on the keys suggested perfect fidelity: the parts of your piano would move exactly how they did when Rachmaninoff or Paderewski played. Would this recording, played on an actual piano, count as “live?” 2

Player Pianos were also an important instrument for the American composer Conlon Nancarrow, who wrote extremely complex pieces using punched paper rolls, pushing the machines to their limits and beyond the limits of the human ear. Sound and performance artist Michael Yuen tells me that:

Nancarrow had to specially alter his player pianos. they were fitted with vacuums to suck air out quicker. There are times when his music calls for 25 notes per sec. That’s a note every 40 msec. The ear can only hear a difference at best at 20msec. Without the turbo charged pianolas, the mechanism pulling the air couldn’t move the hammers and notes go missing. 3

I’d be interested to find out if these machines were popular in China, and if musical production of this sort has influenced the current generations of artists and musicians?

  1. Intermorphic (2010) Mixtikl 2: THE Generative Music & Loop Mixing System. [Online]. Available from: [Accessed 5 March 2010].
  2. Seaver, Nick (2010) OLD MEDIA: INTERACTIVITY AND MECHANICAL MUSIC. noise for airports. Weblog. [Online] Available from: [Accessed 5 March 2010].
  3. Yuen, Michael ( (10 February 2010) Re: generative v. hand-crafted. Email to: Sanderson, Edward (