I have just a few more comments on this subject.  Again, thanks to
everyone for helping.

Robbie added a footnote to my last post which gives a valuable insight.
The data file he eludes to in his post was not the final version.  I
can't find the first master roll.  Robbie's data came from that roll.

After proof copies were made from the first master roll, I made a few
corrections.  Those corrections were physically made to the proof copy.
Once I was satisfied, I phoned Richard Tonnesen and gave him the exact
location of each new perforation.  When we were certain he understood
exactly my corrections, he entered the new data directly in his
perforating file.

The date of those corrections was 6/28/1992.  For all of us, that date
predates the use of the Internet as we know it today.

In his phone call to me Ed Gaida said the final file date for the roll
was 6/28/1992.  When you watch the finished YouTube, you will see that
date on the proof roll as the video begins.  Later in the video, those
slight changes can be seen as they roll passes over the tracker bar.
Don't expect to be able to read those changes as the blue colored
marker is faint.  The pencil marks are the precise measurements from
known perforations to the site of the  changes.

If we were doing this today, I would merely make a change in the data
file and send it to Ed Gaida on the Internet.

Concerning Robbie's remark about the elapsed time of the roll being
5 minutes 54 seconds, that is exactly right if the roll is played as
a MIDI file.  But within the piano, the take-up spool effective diameter
increases during playing, so there is a slight acceleration in tempo.
Not much, but enough to reduce the playing time about 10-15 seconds or
so.  Then if the tempo is set to 67, a further reduction in time occurs
of about 11 seconds.  My finished recording has an elapsed time of 5
minutes and 25 seconds.  So I think the tempo is reasonably close to
ideal.

The roll perforating community is really a great group of dedicated and
talented individuals.  Now with computers and optical scanners, their
preservation work is much more than a utility.  It is a science and
also it is art.

Bob Taylor
Missouri