Don Winter wrote:

> Doug Henderson's article touches on my own mild frustration.  The
> pneumatic pianos have the drawbacks of irregular music and short
> length, as well as historical accuracy of the recording.  It seems
> to me that solenoid pianos should run circles around the older guys.

There are lots of different actuators in industry today, including
pneumatics that are far improved from those of the 1920's, and
electromagnetic actuators of every conceivable variety.  Controls have
similarly improved over the last few years as new power electronic
devices have been developed.

> In order to build a fantasy piano, one would have to understand pianos
> themselves, music itself, pneumatic pianos, electronics, computer
> hardware, computer software, MIDI possibly, and have the skills to
> play the piano, woodworking, metal working and so forth.  Economics,
> geography and agreement on the project at hand mean a combined effort
> is not likely going to happen either.

Well, a lot of these requirements are satisfied if you start with a
piano in good working order, which players/reproducers generally did.

> My fantasy piano would be a 9' grand, with a CD drive digitally driving
> a set of solenoids.  The music would be created as Doug does, on a
> computer, note by note, controlling the expression, saving the file
> to the CD and tweaking the recording until it did indeed sound like
> Gershwin, or Waller, and so forth.  Over time, a nice library of CDs
> would become available because others would sit down at the computer
> keyboard and punch in their favorite tunes and arrangements.

> The work done thus far with solenoid pianos is difficult to understand,
> much of it either not explained in simple terms, or possibly kept as a
> trade secrets.  The technology evolved from existing technology rather
> than ground up new stuff.  The manufacturers market the items through
> dealers, with prices unpublished so that the hobbyist who might install
> one himself is discouraged.

My guess is that it's all a lot simpler than they're letting on.  The
solenoids would have to be stock items, as would the control electronics.
All the rest is software, which is of course significant but hardly in
the realm of mystery.  Industrial machinery performs mechanical tasks
that must be at least as delicate as the pressing of a piano key.  I'm
afraid that piano manufacturers still market their stuff like it was
1938, with all the secret ingredients and processes.

> I guess after nearly 40 years around pianos and computers I have come
> to the understanding this should be the ideal. I have a friend who just
> bought a set of speakers for $7000.  His tone-arm without the cartridge
> is $2500 and he has a $2000 cartridge enroute.  I am to hear this
> system in a few weeks.  Using vinyl recordings, I already have the
> mindset that this system will not perform to my satisfaction, given the
> expense.  This whole "business" of recorded music must have a mindset
> that encompasses, ego, status, traditional peer group approval and a
> whole host of characteristics that have nothing to do with ideal sound
> quality.

About the best thing I can say about high-end audio is that it's generally
not a good idea to argue with someone's religious beliefs.

Insofar as the sound quality goes, it's vastly subjective.  If you believe
that it's poor, then that's what you'll perceive.  I think it was Leonard
Feldman, the guy who first figured out how to market high-fidelity sound as
if it was the salvation of mankind, who came up with the term
'psychoacoustics' to explain why a skilled audio salesman can sell just
about anything to anyone.

As it stands, we can now reproduce music such that, from a mathematical
standpoint, nobody will be able to tell the difference (though we heard this
claim in the 1940's, too.)  The variables -- room acoustics, background
noise
in the listening room, and conditions at the recording studio -- are out of
the control of the designers of the reproduction chain.

> I am realistic enough to know that this is not going to happen, that
> the demand for such an instrument is far too low to be viable but it
> is fun to dream about it!

You can probably do more than that.  Have a look at industrial catalogs that
feature small actuators and, perhaps, controllers that can be connected to a
PC.  That would get the project started.  I'm certain that others on MMD
have done similar investigations.

Mark Kinsler - who often speculated on doing this himself.
Lancaster, OH USA
http://www.frognet.net/~kinsler