James:

I am pleased to make your acquaintance by e-mail.  It is always nice to meet
someone who shares our mutual interest in rolls and scanning.  This note is
in response to your questions in the Automatic Music Digest dated 96.03.06.

Cakewalk is entirely adequate for examining selected portions of the scanned
roll image in detail.  It displays only a small part of the roll on the
screen; this is a disadvantage if you want to see an overview, but it is very
useful for examining particular regions for the purpose of comparing them
with the paper roll.  You cannot examine the entire roll by eye in any event;
there are approximately 1,000,000 punch positions in the roll, so you will
have to be satisfied with spot-checking individual areas that you find to be
of special interest.  If you want to examine the entire roll in overview,
have Richard Tonnesen perforate a new copy of the roll for you from
68283B.MID and superimpose this new roll over an original roll made in the
1920s.  The paper that Richard uses is translucent and lends itself very
nicely to this sort of visual comparison.

It is a mistake to try to play the file 68283B.MID.  Although the file bears
the ".MID" extension, it is not a MIDI file.  Rather, it is a file image of
the (recreated) master roll  used to perforate Ampico Roll No. 68283.  More
correctly, it is a translation of the roll image from my proprietary roll
image data format into MIDI file format for interchange purposes.  To hear
the performance, you need a program that models the behavior of Ampico
instruments.  For a (translated) true MIDI file suitable for playing, I
suggest you get in touch with Robbie Rhodes (rrhodes@foxtail.com) who will
most likely be willing to perform the translation for you.  (Robbie:  My
thanks in advance for helping James and others who want to hear this
performance.)

To answer your final question, note that Ampico Roll No. 68283 occupies
approximately 23 kilobytes when compressed by Jody Kravitz for dissemination.
Taking this size as an average (the roll is about 10 m long, and therefore
typical of Ampico rolls), we find that 150 megabytes of storage would contain
the entire Ampico library of  6000 rolls.  I assume that if this body of work
were available, somewhere there would be someone who would be willing to copy
the entire library onto the "modern" storage medium every 20 years or so.
(Note that currently a single CD-ROM has a capacity of more than 600
megabytes, sufficient to contain all Ampico, Duo-Art, and Welte-Mignon
rolls).  Such
copying would take only about an hour.  I am not concerned about long-term
data loss nearly as much as I am about the deterioration and possible
destruction of the original paper rolls, which appear to be near the end of
their useful life.

Finally, let me suggest that since we now have the ability to make exact
copies of the rolls, an interested party (not me!) could perforate new rolls
from the (recreated) master rolls and start the aging cycle anew.  The rolls
would not necessarily have to be made from paper, as in the 1920s; a more
suitable substrate might be mylar or some other long-lasting synthetic
material.

Sincerely Yours,

Wayne Stahnke