When I play my 1911 Steinway upright Themodist piano for new visitors
and friends who may have often heard it before, they often ask me how
the rolls themselves came into being.  That is a different question
from how the perforations are mechanically punched in the paper, how
these holes interact with those on the tracker bar, and how the genies
behind the tracker bar are awakened to activate the notes, the dampers
and so forth.

My friends mean, "How were the actions of the pianist's hands and feet
converted into holes on rolls?  How did performances get turned into
rolls?"

I reply with something based on thirty some years of scattered reading,
but I don't really know the answers.  Nor am I going to do research
right now, because I'm not sure where to look or whom to believe.

When friends ask whether the original "recording" sessions were
analogous to today's electronic taping sessions, I opine that they were,
or at least some were, sort of.  I tell them I believe, in many cases,
there was something called a "marking piano" that made notations on
rolls that later were translated into perforations.

I say also that the major roll manufacturers had gifted editors who
helped to create, and sometimes to improve upon, those markings and
perforations.  But just how much the editors did, or what they did,
I cannot say for sure.  I get into even murkier territory when I say
that I've heard that some expert, like Frank Milne maybe, was even
able to create a roll with no need for a piano until playback time.
(Is that true?)

I say too (and here I'm on really shaky ground) there were a few
devices that directly perforated the rolls to match the performance
(in a manner approaching today's "real time").  I least I think that's
true.  I say that, on one visit to Keystone Music, Richard Groman showed
me the very piano (a ghastly green, I think) on which J. Lawrence Cook
made many of his rolls.  (But I don't know whether that was a marking
piano or not.  I do know Richard said it had mechanical couplers that
let the pianist to create more notes than any ten fingers could play
in a given moment.)

In short, there is a lot I can only guess at regarding what went on
between when Mr. Rubinstein sat down at the piano and when the people
at Ampico or Duo-Art played the "recorded" result back to him.  (Not
only the transfer of "his" notes to paper, but all of those nuances
of how hard he struck the notes, how long he held each, how he used 
at least two of the pedals, and perhaps a lot more.)   And I guess it
might be simpler to think about the data on 88-note rolls, but it's
not as much fun.

I'm confident that many readers of the MMD actually know the answers
to these questions.  And I know a few books and articles have been
written on this subject.  What I'm hoping to spark is a discussion of
which available accounts of the roll development process to believe and
reliably learn from.  Not only would I like to be able to answer my
friends' questions: I'd like to know the answers myself.

Paul Murphy

 [ The 'drafting board' method came first, and it continues to be
 [ used to this day because it's the cheapest way to get the job done.
 [ J.L. Cook's piano is a variant on this method.  The pencil marking
 [ piano (analogous to today's live MIDI recording) was used primarily
 [ to record classical music performed by big name performing pianists,
 [ an expensive and time-consuming process that wasn't justified for
 [ rolls of popular music.  The "real time" perforating machines were
 [ mostly experimental and never displaced the other, more economical
 [ methods.  -- Robbie