After my last posting, I got some emails asking about the quality of our
new rolls and about our roll scanning process.

Maybe I need to explain the roll copying process and our capabilities in
a little more detail.

Being a collector and restorer of mechanical instruments for about
thirty years, I have had the chance to learn a lot and gain quite a bit
of experience which was very helpful when I started to design my roll
cutting equipment about fourteen years ago.  I was aware of the many
different roll layouts, and the very first roll I wanted to cut was a
Welte Philharmonic roll.  I thought that if the new cutting machine was
accurate enough to reproduce a 150-track Philharmonic roll with its
close spacing, I would have no problem with any other kind of roll
in the future.

The first step was to read the roll accurately.  Wolfgang Haas and I had
already build a pneumatic roll reader and, after dealing with a lot
of problems, we were finally able to get a more or less accurate MIDI
file from the roll.

By coincidence, I got at that time into contact with Prof. Kieseyer of
the Institute Of Digital Data Processing at Wuppertal University.  He
was very impressed with our experiments and enthusiastic about their
possibilities.  We had many discussions with Prof. Kieseyer and finally
came to the conclusion that an optical roll scanner would be the best
solution.

In the end Prof. Kieseyer offered to us the opportunity to involve his
students in the development of a roll-scanning unit and the essential
software.  In 1994 that was a unique task!  Neither software nor
hardware for that purpose was available on the market.

As a first step I had to come up with a detailed plan and layout for all
the parameters that would be necessary to read any kind of roll.  For
the next four months I traveled almost every other day to Wuppertal to
work on the development of an optical scanner.  I had to build the
mechanical parts for the roll feeder plus the camera unit.  The software
and hardware for the optical reader was than developed at the Institute
according to my specifications, using the latest available Sony CCD
technology.  The project eventually became a degree dissertation for one
of the Wuppertal students.

As a result, we have had since 1994 an extremely accurate optical roll
scanner plus very comprehensive and versatile software to transpose any
roll into an accurate MIDI file in any degree of resolution -- if
necessary up to a thousandth part of a millimeter.

Many additional useful features were built into the software as well.
Some examples: the software allows us to descramble organ rolls, to view
the music on the screen, and to make repairs to a broken roll.  We can
audition the music as well before cutting a roll.  The software allows
us to read any media, regardless length, number of tracks, or uneven
spacing.

There is an option to read and cut a roll hole-by-hole, which we use
with some rolls, for example Hupfeld Helios.  We can read the beginning
and the end of each note as well, allowing custom spacing for the new
roll, which has proven to be very useful in practice, especially if the
original roll was badly cut with very uneven spacing.

The second step was the development of a universal and very accurate
roll cutting machine to reproduce any kind of roll.  In 1994 standard
printers used paper with side perforations to turn out very accurate
prints.  I liked this idea, thinking that this might be a good way of
safely guiding a roll during the perforation process.  I discussed the
idea with my friend Ernst Schwab, who volunteered to write a custom
software program for this purpose.  Ernst Schwab turned out to be the
perfect partner; he immediately understood what I needed and was able to
design the software to exactly reproduce any kind of roll.  His software
allows us to do any kind of modification to any kind of roll.  Then we
built a sturdy mechanical roll perforator which uses 4 chains to drive a
paper roll with side perforations, eliminating any slack in the paper.

I would sometimes impress visitors by stopping a roll in progress and
telling the computer to go back to the beginning and start the roll
again.  The perforator would then cut exactly into the same holes which
were already cut!

After completing the cutting of a roll, each copy is then fed through a
special machine that cuts off the side perforations.

This methodology has proven to be the right way to go.  The results have
been very satisfying.  Using a single cutting head slows down the
cutting process a bit, but allows us to cut every kind of roll with any
kind of spacing.  Even rolls with mixed and uneven spacing are easy to
do.  The equipment allows us to transfer music from one system to
another as well.  As a result, we can now reproduce any type of roll
faithful to the original.

There is another often ignored, but extremely important, consideration
in cutting new rolls: paper quality.  There is no kind of paper that is
100% unaffected by humidity.  Any kind of paper tends to shrink or swell
more or less due to the humidity in its environment.  Many collectors
know what I am talking about.  I acquired many rolls in the past which
had been stored under bad conditions, either too damp or too dry,
resulting in a roll which would no longer track properly.

Being aware of that problem, we did a lot of investigation.  It was
quite a task to find a high-quality paper to suit our needs.  After
bugging several paper manufacturers in Germany, we finally found a
source for a satisfactory high-quality paper.  This company is able to
deliver a kind of paper that is very resistant to humidity changes.  It
is more expensive than other paper, but we decided in 1994 to use it,
and again it turned out to be the right decision.  So we are able to
reduce paper expansion and contraction in the rolls we make to a
minimum.

I hope this explanation is helpful in understanding how we make music
rolls.

Best regards,
Thomas Jansen
http://www.maesto.com