Wayne Stahnke states that his goal is "exact scanning of the roll".
I believe that the ultimate goal is more encompassing:

   "The goal is to transcribe _all_ the data of the music roll
    into a binary file, such that an exact replica can be made."

In order to "demonstrate that this goal can be achieved", one can
produce (perforate) a music roll from the binary file, which may
be compared with the original roll.  Indeed, if this new "replica"
is identical to the original, hole-for-hole and scallop-for-scallop,
then that is _conclusive and complete_ proof that the copying
process has recovered all of the data from the original roll.
(Obviously, copying the printed lyrics and the fancy artwork are
separate processes.)

If an exact replica of an original music roll can be made, then
if follows logically that all the data of the Master Roll has
been recovered, and one can truly declare: "This computer file is
an exact copy of the Master Roll, and production rolls produced
from this file will be identical in all respects to production
rolls made decades ago."

Richard Tonnesen is equipped to produce music rolls from Stahnke's
Master Roll files.  If anyone has a supply of genuine Ampico
paper from 1920, a "counterfeit roll" can be made which cannot
be distinguished from an original.

You can't do any better than that!

-- Robbie Rhodes