Karl Petersen's description of a sideways cylindrical punch that would
more cleanly append to continuous punched lines is a nice plus to the
cupped punch theory.  The punch I saw looked more like a cupped potato
chip - but K.P.'s method (maybe I picture a tilted cupped potato chip?)
will help yield a physically exact copy as he describes.  This got me
thinking about roll bridging.

Some rolls don't employ any bridging.  These are rolls that seem to be
torn the most, or when they require moderate tracking, buckle easily.
I guess that most roll manufacturers inserted bridging to produce a
mechanically stable roll with no effect to the music.  Inspect a
Welte/DeLUXE roll...  their note lines are punched full round holes
minutely spaced, which intellectually seems to be the most physically
stable.  Other rolls seem to use bridging after an inch or so on a long
note, presumably to give the pneumatic the fullest atmosphere for
attack, then comes the bridging that provides lateral strength yet
enough atmosphere to sustain it's state.  Might it be true that perhaps
on some reproducing units, where tolerances are tight, changing a
copy's bridging pattern will alter how the music's played? If true,
you'd better stick to the original patterns.

I guess a Personal Perforator might have two software settings.  One
would be for an exact duplication where the use of bridging is
critical, or just because you want an exact copy, thank you very much.
The other mode would be for an "enhanced" duplication that employs only
close full punched round holes; useful for enhanced physical stability
and for copying MIDI files where there exists no bridging code.

This stuff's fun to think about!

- K a r l   B.   E l l i s o n
  Ashland, Massachusetts  U.S.A.
  KBEllison@aol.com
  http://members.aol.com/kbellison/kbe.html