Bruce Clark's posting on Tempo Problems in Spliced Medley Rolls (980421)
and more recently Harvey Chao's comments on a mis-punched QRS roll
(980525) and finally Tim Baxter's Maximum Notes-Per-Second on Piano Roll
(980528) have spurred me into posing the following question for the wise
men of the MMD.

Is it reasonable to expect a player mechanism to respond reliably to a
single round perforation on a roll when the expression level is low (P or
PP) ?

I've just taken delivery of the Recordo three-roll set of Schumann Sonata
Op. 5, played by Beryl Rubinstein, or Rubenstein. (It depends which roll
label you look at; he recorded several pieces for QRS).  Anyway, the first
movement of the sonata has many staccato chords and these are coded onto
the roll as single round perforations.  My Gulbransen Recordo upright can
cope with these when the music is loud but it certainly doesn't when it is
soft.

These rolls come from Rock Soup (Ginny and Bob Billings), and I know that
they are very fussy about the quality of their recuts.  A friend of mine
has an original of the first roll in the Schumann set and I have borrowed
it to do a hole-for hole comparison of at least the first few feet of the
original and the copy.

Actually I used metric measurements because that's what I'm used to.
Anyway, the first 2.378 metres of music on the original roll occupies
2.387 m on the copy.  That's 0.38% longer and who's going to quibble
about that?  Humidity changes will have a much greater effect and
modern paper will probably respond less to humidity than old paper.

Why 2.378 m ?  That's the usable length of our table tennis table.

The chain perforation patterns on the two rolls are quite different but I
much prefer the new pattern.  There are many more bridges along the long
perforations, making the roll a lot stronger.

But it's the length of the perforations for staccato notes on the recut
that has me worried.  On the original the holes are longer than they are
wide: on the recut they are a single round hole.  I've measured the length
of these slots a couple of times with a millimeter scale and a magnifying
glass.  Typical results are:  Old length  1.8 mm   New length  1.55 mm
          "       1.7 mm       "       1.4  mm

This indicates my method is not very precise but there's a clear
difference.  Bob Billings says he is pretty sure that the Tonnesen's
punch is 1.8 mm in diameter.

The ports on my Gulbransen's tracker bar are 2.2 mm wide and 1.6 mm high.
I think I recall reading somewhere that a tracker bar hole needs at least
1/7th of its area uncovered before the airflow into the port exceeds the
leakage rate through the bleed, allowing the pouch to lift.  Assuming
this is correct and using 1.8 mm for the hole dia., I worked out how
long a port will be uncovered by at least 1/7th when the roll is
travelling at speed 70, the recommended speed for this roll.  It is
70 milliseconds, or 0.07 seconds.  That doesn't sound very long to me;
is it unreasonably short or not?

Does the old roll play better?  On the Gulbransen it does seem a little
better, but not a lot.  Both rolls respond significantly better when
pedalled through on my Aeolian upright, so some of the problem may be the
bleed hole size in the Gulbransen.  I must confess that I have one of John
Farrell's rolls that I can't play on my Aeolian, because the bridges in the
chain perforations are rather wide and the player treats each perf as a
separate note.  The tremolo effects are a bit much!  This roll plays well
on the Gulbransen.  (It's Bethena Waltz by Scott Joplin - a great roll by
the way, John, if you're reading this.)

Finally I got out my British Aeolian Themodist version of the music and
looked at the holes on that. The shortest perf is 2.2 mm long and is
obviously a double punch.  No single punches on this roll although I think
I have seen them on Aeolian rolls.

Finally I must make it clear that I'm not criticizing the Billings in any
way.  All we Recordo fans should be very grateful to them for becoming a
source of high-quality recuts.

John Phillips in Hobart, Tasmania.