THE ROLLS AS MIDI FILES

The paper used to make Aeolian pipe organ rolls is not acid free.  As the paper ages, it becomes brittle.  Since player rolls are "read" in a mechanical process and are spooled back and forth between two "cores", the paper is subjected to flexing every time the roll is played.  Also as the paper ages, it warps unevenly as it absorbs moisture from the air.  The warped paper during play makes it difficult to keep necessary alignment factors within operational limits.  The longevity of the paper roll appears to be reaching its limit.

Careful storage of the rolls will aid in longevity.  Another way to preserve rolls, is not to use them.  This is a dilemma.  So the best solution is to copy the data from the roll, safely store the roll, and use the copied data to play the organ or make new rolls.  

To copy the data correctly is no small task.  The data collected from the roll can be used in two general ways.  When used as performance data, the data must reproduce the same results as the roll moving across the tracker bar.  When used as perforation data,  the data must drive a perforator in a manner to exactly duplicate itself.  Casual treatment  of these two distinct processes in the past has resulted in some copied rolls, mostly for piano, that contain huge inaccuracies.

SPEED BUILD UP

When playing an organ roll, the speed that the paper moves across the tracker bar slowly increases as the take up spool becomes constantly  larger due to paper build up.  If the data collected from the roll contains this characteristic of slowly increasing speed, then the data will play the organ at correct tempo. If performance data is used to make a new roll,  the new roll will contain this acceleration, thus yielding a double speed increase when the new roll is played. 
CORRECT "DURATION"  FOR EACH HOLE

As single dot perforation crosses the tracker bar opening, an "on" event is generated.  The duration of the "on" event is a factor of roll speed, the size of the paper hole, and the size of the tracker bar hole.  In the case of the Aeolian, the tracker bar hole is large relative to the paper hole.  This means that the duration generated by a single hole is longer than the time required for the single hole to pass a fixed point.  In fact, it is about twice as long.  The collected data containing this characteristic (performance data) cannot be used, without correction,  to make a duplicate roll as the new generation roll will have an elongated perforation.   Failure to compensate will result in the complete loss of all staccato notes and other errors.

The goal of this project was to create performance data files.   A complete original organ roll player was converted to to serve as the data collector..  Powered by an electric motor with governor, the paper roll mechanism moves very slowly at speed tempo 20.  The electric contacts triggered by the holes in the paper generate signals that go from a MIDI processor into a computer file.  To aid in monitoring the accuracy of the process, an electronic organ module generates audio sound of the data being recorded.

On screen editing is used to clean up the recording if errant notes occur due to mistracking or roll damage.  The correct intended tempo of the composition is restored, and the file is saved.  Further editing is accomplished to take full advantage of the organ resources and its MIDI interface.  The final result is a data file that drives the organ exactly as if the roll were playing the organ.

There seems to be no standard, at this time, regarding various techniques of copying paper music rolls.
While optical scanning and pneumatic reading can be accomplished with great accuracy, if the  data is undefined regarding its  specific  characteristics, the data cannot be properly altered for various uses.  

A detailed discussion on data collection and accuracy is here.

What about the printed instructions on the 116 note roll?   That is another story.  Click here for more.

home