Joe Orens asked, "How much difference does the size of the "Themodist"
holes make on Duo-Art recut rolls?"

The "snakebite" has the effect of switching the appropriate half of
the stack from the "accompaniment" suction level to the "theme" suction
level, which is usually (but not always) higher.  (The exception is
where snakebites aren't needed for the reproduction but are provided
to add accenting capability for people using the manual controls,
or on "half-Duo-Arts", pedalling the rolls using Pianola controls.)

This switch-over is not a crisp electronic-style "square wave".  The
theme blip resulting from the shortest possible snakebite, after it
has been relayed through the theme primary and secondary valves and
has vacuum-cleaned the great cathedral expanse of the stack, is a
quite sharp but nevertheless rounded pulse.  I once measured it using
a telephone microphone insert (!) and it looks something like this
(view with a fixed-pitch font, e.g., Courier) :-


Vertical = suction,  horizontal = time

      |- 0.1 sec -|   (assumed roll Tempo = 70)


        .      <-- Theme power
       : :
       :  .    This slope steepens if there are a lot of notes open.
       :   .
........    ..............   <--  Accompaniment power

       ^-- snakebite opens "theme" port


Notice that nothing at all happens until most of the snakebite has
passed the theme port, so snakebites are normally placed one row
(punch advance) ahead of the start of the note to be accented, to get
the blip where it's needed.

The relatively slow return to accompaniment power explains why as a
rule Duo-Art theme notes are delayed rather than advanced relative to
the nominally simultaneous accompaniment notes.  If you advance the
theme note, some of the blip carries over to the start of the
accompaniment notes, and while this can be (and is) allowed for in
their power coding, it lends a degree of uncertainty to the
performance of the instrument, which then depends on consistency of
Themodist valve transition speed, which otherwise isn't critical.
If you delay it, the accompaniment pneumatics are well on the way to
closed before the blip occurs and it has no effect on them.

The next thing to be said is that a single snakebite doesn't quite
raise the stack to the full theme suction level, although it nearly
does.  A double one does and depending on the number of notes open, an
oversize one may do.  Depending on the nature of the music, this might
make a noticeable difference, perhaps as much as two Duo-Art steps
where there's a big difference between the two power levels.

So for really good reproduction, if you're not going to recode the
rolls, you really must use the original size of snakebite.  Their tiny
size was precisely to be able to differentiate between theme and
accompaniment notes in the shortest possible time and instruments
that don't are simply going to sound different - although on some
rolls the difference won't be detectable.  Two holes were provided to
achieve the snappiest primary valve operation available (they also
have the virtue of ensuring that "themodising" continues to work if
the roll is slightly off track).  On 65-note rolls, for which the
snakebite system was originally designed, they're the same size as on
88-note rolls, where the note holes are 50% fatter.

Nevertheless you can get away with limited murder.  I remember Paul
Johnson asking me, when he decided to issue Duo-Art recuts, whether I
thought the snakebites produced by the Tonnesen perforator, which are
the same size as note holes, were going to be acceptable.  I think he
was a bit put out when I insisted that he should at least advance the
snakebites which related to advanced themed notes (advanced relative
to the accompaniment note starts, that is) and delay those which
related to late themed notes, to produce the least added influence on
the accompaniment playing strength.  (Of course even this doesn't work
in chords where a single snakebite is trying to pick out one note in
a spread.)

I haven't examined the resulting rolls with a magnifying glass, but I
have noticed that Denis Hall, an acute London critic of recut Duo-Art
rolls, having heard some of my pink-label Piano Roll Shop recuts, has
been getting in his own supply in suspiciously large quantity.

The British "Play Time" recuts, produced by Steve Cox of Laguna Rolls
using a very fine (and very temperamental) pneumatic roll reader, are
punched with original-size snakebites and while, being very accurate,
do not quite achieve the Holy Grail of "punch-for-punch" copying.
But that's another story ...

Dan Wilson
London


 [ Editor's note:
 [
 [ The ubiquitous carbon-button microphone was used in telephone
 [ handsets throughout the world until recent years, when it was
 [ replaced by the electret condenser mic.  The carbon mic had
 [ a characteristic mild audio distortion, but it was virtually
 [ indestructible !
 [
 [ Dan's application of the carbon mic as a pressure transducer is
 [ ingenious: the current flowing through the carbon button varies
 [ more-or-less linearly in proportion to the air pressure (or vacuum).
 [ Dan, did you power the circuit like the traditional telephone,
 [ using a big "No. 6 Dry Cell"?  :-)
 [
 [ -- Robbie