For the first time, the Stanford early-recordings symposium is being held only a year after the last one. For the first time, also, it's strictly only two days long (it was first announced for three), which may be because they didn't allow time to fully recharge the tank. They've also moved it back to the tiny department auditorium where the first two were held, because there's hardly enough people attending to even warm that room up.
This year's entirely concerns piano rolls, Stanford having become a major research center on them last year when they bought a ginormous collection of them (now inventoried at 7450 rolls) plus about a dozen instruments. Piano rolls, one of the speakers said, are like portraits, where recordings are like photographs: technically, the latter are a better likeness, but especially in the early days of each, the former often captures the spirit better.
Player pianos, it turns out, are of two kinds, fostering two approaches to making the piano rolls:
1) Reproducing pianos: full-scale pianos, usually uprights but could be grands, with a roll-playing mechanism inside. If you've seen one, it's probably this kind. The mechanism is usually run by electricity (though there were water-powered ones). The rolls are typically interpretive performances by great pianists, and the roll controls the dynamics and the pedal as well as the keys.
2) Pianolas. "Pianola" is actually the brand name, and the company made reproducing pianos as well, but the term generally means these. The generic term is "push-up". A pianola is essentially a robot that appears to play the piano, though it actually requires considerable human intervention, like the Great and Terrible Oz. It's a cabinet with a roll-player mechanism inside and 88 mechanical fingers sticking out the back. You push it up to a piano keyboard (hence the generic name) and the mechanical fingers play the keys. The difference is that it's not run on electricity; it needs human feet pump-action to work. This controls volume; tempo and, confusingly, pedaling, are controlled by hand-levers.
For this reason the rolls were usually mechanically-exact punchings out of the score. It was up to the human operator to provide the artistic content by manipulation of the pedals and levers. This can be difficult, but the results can be rewarding. At today's session, the world's greatest living pianola player (who is also in the running for the world's longest beard and the world's baldest head) demonstrated the difference between doing it well and badly. Most CD transfers you'll hear of these rolls? Badly; very badly indeed.
Then his colleague told us about reproducing-piano piano rolls of the solo parts of concertos, and of how tough it is to accompany them. (We'll be hearing one at the concert tomorrow night.) The part I liked is how the rolls have special punches that make them stop running when it reaches a part where the orchestra is playing without the piano. Then, when the piano is to re-enter, the conductor presses a silent electric buzzer that starts it up again, one hopes in time.
After lunch, we turned to what can be done with piano rolls in the modern age. One guy who obviously has way too much time on his hands has been scanning and digitizing multiple copies of the same piano roll and measuring the microscopic differences between them. Another was an academic composer who found himself gobsmacked by the idea that a player piano could play, not just the great masterpieces of piano literature, but music physically impossible for a human to play. So thrilled was he by this that he went out and wrote such a piece, which we'll also hear tomorrow. It puzzled me that he then alluded to Conlon Nancarrow, because Nancarrow had the same idea 60 years ago, so if you've heard of him, it shouldn't seem like such a novel mind-bender.