The earliest efforts to copy original paper rolls were
based on the desire to make new copies of the original roll. Almost
immediately it was realized that the task is more complex than it appears.
Before computers were used, the "one for one" copies rarely could
duplicate a roll with reliable accuracy. The problem was basically one
of synchronization and calibration. Without getting into all the reasons,
the resulting roll would seldom match when it was laid on top of the
original. Those who recognized the problem, and who were diligent,
were able to incorporate adjustments to bring the differences into acceptable
tolerance. The important factor in these early efforts was strong, reliable
feedback. The feedback; comparing the new roll to the old, was immediately
available and reliable.
With the advent of computers, rolls were copied and the data stored, to be used at a later date. The data was also distributed to users who used the data in several ways. There are a variety of techniques of collecting the roll data. The speed of the paper passing the collecting device may be constant or may slowly increase as the paper builds up on a take up spool. The hole in the paper may be sensed by optical scanners or by pneumatic means. After the data is collected, it may be altered by various computer programs to accurately control perforators or electronically control music machines. The intended use of the data can be as either performance data or perforation data, both may be archival quality.
The challenge now becomes identifying the exact characteristics of the reading apparatus. If full documentation is lacking, the data cannot be applied by the end user with full accuracy, and certainly cannot be classified as "Archival". The feedback that was used in first generation copy efforts, requires side by side comparison, and that is not always available. In fact, there is no feedback at all for performance data (except for subjective listening). For archival data, the collection characteristics must be completely defined. In defining collection characteristics, a full understanding of the Bleed Balance Point and other factors must be understood.
Bleed Balance Point refers to the threshold at which an "on" event is first generated as a perforation progresses across the tracker bar opening. In most pneumatic player systems, small bleeds are used to equalize the chamber under a pouch. As the hole of the tracker bar is opened by the note sheet (roll), the initial opening must overcome the bleed, and flow restrictions in the system, before the pouch can trigger the "on" event. In properly designed systems, the bleed balance point is reached well before the note sheet hole reaches 50% open. It appears as if the Aeolian has a bleed balance point at about 25% of linear travel across the tracker opening.
In the case of Aeolian Pipe Organ Rolls, the chance for scanning error is significant. This is because unlike most 88 note piano rolls, the pipe organ roll has holes that are small when compared to the opening in the tracker bar. In the illustration below, two positions of a single perforation hole are shown. The perforation moves from "note first on" down to "note first off".
1. Roll paper speed
2. Tracker hole size
3. Roll hole size
4. Bleed Balance Point
5. Contact box adjustment (Aeolian contact boxes have adjusting screws)
In this illustration, if the roll speed is tempo 50 (5 feet per minute), the proper time of the "on" duration is about .0865 seconds. Ignoring the tracker hole size (timing the .043 single perforation past a fixed point) truncates the event with the duration becoming 50% of designed value. Including the entire overlap time in the "on" calculation results in an increase to 175% of the designed value. This demonstrates the possible wide error range .
If an Aeolian Pipe Organ roll is read pneumatically by standard Aeolian equipment and the output signal fed into a MIDI file, then that file will possess all the characteristics required of an accurate performance file. If the same roll is optically scanned, the "on duration" of each note must equal the same duration of the pneumatically read notes. For that to happen, the optical scanned roll data must be processed and calibrated. The calibration can be through construction design of the scanner and/or by computer manipulation of the data. Optically scanned data must contain documentation showing their relationship to normal pneumatically read rolls. Simply stated, most optical scans are just pictures of the roll and playing a picture requires special programs. The challenge is getting the program right.
Currently, there is a proliferation of optical scanners. While most applications of the optical scanner are not as critical as Aeolian Organ Rolls, there is still a lack of standardization which leaves wide open the question, "Which file is of archival quality?"
Documentation of collection characteristics should be included in all data files. One method could be to make each data file self explanatory by including several test events at the beginning of each roll. The same events should also be included at the roll end. Those files can then be scrutinized at a later date and with proper analysis, can be varied as good usable data.
Back to rolls as MIDI files
home
With the advent of computers, rolls were copied and the data stored, to be used at a later date. The data was also distributed to users who used the data in several ways. There are a variety of techniques of collecting the roll data. The speed of the paper passing the collecting device may be constant or may slowly increase as the paper builds up on a take up spool. The hole in the paper may be sensed by optical scanners or by pneumatic means. After the data is collected, it may be altered by various computer programs to accurately control perforators or electronically control music machines. The intended use of the data can be as either performance data or perforation data, both may be archival quality.
The challenge now becomes identifying the exact characteristics of the reading apparatus. If full documentation is lacking, the data cannot be applied by the end user with full accuracy, and certainly cannot be classified as "Archival". The feedback that was used in first generation copy efforts, requires side by side comparison, and that is not always available. In fact, there is no feedback at all for performance data (except for subjective listening). For archival data, the collection characteristics must be completely defined. In defining collection characteristics, a full understanding of the Bleed Balance Point and other factors must be understood.
Bleed Balance Point refers to the threshold at which an "on" event is first generated as a perforation progresses across the tracker bar opening. In most pneumatic player systems, small bleeds are used to equalize the chamber under a pouch. As the hole of the tracker bar is opened by the note sheet (roll), the initial opening must overcome the bleed, and flow restrictions in the system, before the pouch can trigger the "on" event. In properly designed systems, the bleed balance point is reached well before the note sheet hole reaches 50% open. It appears as if the Aeolian has a bleed balance point at about 25% of linear travel across the tracker opening.
In the case of Aeolian Pipe Organ Rolls, the chance for scanning error is significant. This is because unlike most 88 note piano rolls, the pipe organ roll has holes that are small when compared to the opening in the tracker bar. In the illustration below, two positions of a single perforation hole are shown. The perforation moves from "note first on" down to "note first off".
The purpose of the illustrations is to demonstrate the total time of the
"on" event or single perforation duration. That time is a function of these factors:
1. Roll paper speed
2. Tracker hole size
3. Roll hole size
4. Bleed Balance Point
5. Contact box adjustment (Aeolian contact boxes have adjusting screws)
In this illustration, if the roll speed is tempo 50 (5 feet per minute), the proper time of the "on" duration is about .0865 seconds. Ignoring the tracker hole size (timing the .043 single perforation past a fixed point) truncates the event with the duration becoming 50% of designed value. Including the entire overlap time in the "on" calculation results in an increase to 175% of the designed value. This demonstrates the possible wide error range .
If an Aeolian Pipe Organ roll is read pneumatically by standard Aeolian equipment and the output signal fed into a MIDI file, then that file will possess all the characteristics required of an accurate performance file. If the same roll is optically scanned, the "on duration" of each note must equal the same duration of the pneumatically read notes. For that to happen, the optical scanned roll data must be processed and calibrated. The calibration can be through construction design of the scanner and/or by computer manipulation of the data. Optically scanned data must contain documentation showing their relationship to normal pneumatically read rolls. Simply stated, most optical scans are just pictures of the roll and playing a picture requires special programs. The challenge is getting the program right.
Currently, there is a proliferation of optical scanners. While most applications of the optical scanner are not as critical as Aeolian Organ Rolls, there is still a lack of standardization which leaves wide open the question, "Which file is of archival quality?"
Documentation of collection characteristics should be included in all data files. One method could be to make each data file self explanatory by including several test events at the beginning of each roll. The same events should also be included at the roll end. Those files can then be scrutinized at a later date and with proper analysis, can be varied as good usable data.
Back to rolls as MIDI files
home