Jody,¶
I love your idea of establishing an archive of piano roll data.
And I also think that there should be some discussion, about formats,
procedures etc. before submitting any files.

In my opinion there should be no question about MIDI as "the" desired
platform for this project. The MIDI file should contain the expression
tracks in separate MIDI channels, regardless whether expression has been
interpreted into the notes or not. There should be some accompanying
document that explains the method of generating the MIDI data from
the roll. Methods that have already been mentioned in this list are

-     pneumatic sensing in a more or less original instrument¶
-     optical sensing by means of a modified trackerbar or similar¶
-     optical scanning by a standard scanning device (e.g. Scanjet)¶
-     optical scanning with cameras (CCD line or image)¶
-     ? what else ?

One other property is the "algorithm" that has to be used in some cases
to convert the input above to the notes. I am thinking here about the
influence of the different apertures in optical and pneumatic sensing
and the sometimes not uniform hole length in the trackerbar.
This includes the common problem of removing the (stability) bridges
between adjacent holes in one track, that are ignored by pneumatic
sensing but mostly not by optical methods.

The next item is the algorithm for converting the expression holes
into velocities for the bass and treble notes.

All these things have probably been solved in a dozen places, but I
think these methods should be compared in order to make the data
exchangeable. An interesting comparison could be made if several
submitters provide a file with "their" version of a common roll.

Once we have an approved MIDI file it can be reproduced by all MIDI
compatible devices ranging from PC boards up to digital pianos.

Another interesting discussion that we should start later is the
reversal of the above procedure, which is even more challenging.
The question is how to convert a standard MIDI file into a "roll"
compatible format, that means converting the velocities into the
expression commands. Then this format can be used to punch "new"
rolls or to control midified pneumatic pianos.

Regards

Walter