There appear to be three threads going on at once here, which is
causing a bit of confusion in the discussion.  Perhaps we need to
separate the issues.

   1.  How accurate does the reading of the original need to be?
   2.  How accurate does the copy need to be?
   3.  How do we get there from here?

In the 970819 MMD Julian Dyer expands on the clocking problem, pointing
out the beat problem as it relates to current practice, and Hal Davis
adds to this with his description of different punch advance intervals
in a single multi-tune roll.

The question of accuracy, both in reading and copying, addresses these
issues.  In real life, it is unlikely that any reading or copying process
will exactly reproduce the copy.  However, if we make the sampling
interval small enough, it will be below threshold of perception.  How
this is done is another matter, but today's technology gives us a number
of ways to do it.  Clearly the reading method must be all electronic to
achieve a suitable accuracy (exclusive of the roll advance).

There is another consideration, as Wayne Stahnke points out, and that is
timing of expression.  The Welte system depends critically on timing, the
Ampico B a little less so, and often both Ampico and Duo-Art play some
very tight timing games, sometimes down to the level of a single punch
advance.  In these cases the roll speed and advance intervals become
important.

Wayne and I agree that the goal must ultimately be to capture the
_intent_ of the original master.  The copy may not necessarily do that.
Once we have that we are nearly at Julian's goal of note-for-note copies.
And if we use a high resolution drive for the punch, we can produce
note-for-note copies, within the limits of human perception.

My plan for a punch includes a stepper drive, with 0.001% steps.  This
will allow "close enough" note-for-note copying.  I think it may be
possible to adapt such a system to existing punches, though the interface
would probably require some local smarts in the form of a processor chip
or board.

All the above is the engineer and archivist talking.  As I've said
before, and as Hal Davis said, we should enjoy the positive things in
life (and smell the flowers too).  There are a lot of imperfect things in
this world that give us a lot of pleasure.  Imperfect to a point,
however.....

Bob Billings