This subject has been much discussed.  Let me introduce the following
point: Can a solenoid piano sound like a human performance?  You bet!

There is certainly no technical reason why a solenoid can't be
instructed to mimic keyboard finger pressure and velocity profile.
In the simplest case solenoids are turned on and off with a square
(or if you must, a rectangular) wave.  Solenoid pianos based on this
simplest concept are indeed one trick ponies -- interesting but not
masterful.

It is possible to measure the response of solenoids of varying mass and
elasticity to waveforms of varying temporal shapes of a corresponding
driving voltage profile.  One can correlate the results and place them
in a lookup table corresponding to human finger keyboard touch and the
resulting sonic result.

It is therefore possible to design a solenoid piano that mimics the
mechanics of human piano keyboard touch.

An actual human pianist also listens to his or her playing and
(usually unconsciously) adjusts the touch to compensate for room and
piano acoustics (mostly sound decay time).  A "recording" solenoid
piano based on the above lookup table could be built so that solenoid
playback would produce an identical performance on that same piano in
the same acoustic space.

One could easily add a adjustable decay feature to the solenoid driver
table.  It would be thus generally possible to get the same performance
from different pianos in differing acoustical spaces.  The limitation
here is that the pianos must have similar tone properties.  If the
"recording" piano were a Steinway Model B for example, it is not likely
that an old upright or home spinet would come even close to the
original performance.

Has this kind of thing been done already?  I doubt that the concepts are
unique.  Certainly Boesendorfer has attempted experimental product based
on similar thinking.  Yamaha also has an obvious stake in this market.

Can I buy such a product?  Probably not, and certainly not at a
reasonable price.  It is possible however to get a musician and an
electrical engineer and a piano technician to build such a device using
readily available components.  Any volunteers?

The general thesis of my argument is that solenoid pianos of proper
design can be built to duplicate a human piano performance.  It's not
really rocket science.  There are no fundamental limits to overcome.

Bill Finch

 [ The Boesendorfer SE recording and reproducing piano was offered
 [ for eight or ten years before the product was withdrawn from the
 [ market.  Thereafter the patents were licensed to Yamaha.  The
 [ concept is not just unique, it is proven.  However, the limited
 [ market results in a high sale price.
 [
 [ But the existence of a perfect reproducing piano doesn't mean that
 [ a library of fine music for it exists; I think that's the biggest
 [ drawback the consumer faces.  (More than 4000 different music rolls
 [ were published for the Welte-Mignon reproducing piano!)  -- Robbie