In 991203 MMDigest Craig Brougher expressed much doubt about the
Kastonome accenting system:

> Now I am wondering how it works, of course.  For instance, how was
> it possible to stack accent holes on top of each other in columns and
> yet get the accent for a specific note on the sheet, without also
> accenting all the other notes in the column, which might coincide
> with some notes playing at the exact same time?
>
> I can imagine that the designer would probably stack unlikely notes
> together, but after that all you can say is, the piano would be
> limited to music that didn't play unlikely notes together.

You really need to examine the rolls and the instrument to understand
this.  The only run of notes that would cause a "theme" perforation to
accent the wrong note, as I said in one of my messages, would be a fast
downward arpeggio of notes.  These would also have to be spaced apart
by the "theme" column spacing on the tracker bar.

Douglas Bush has given us, from an actual instrument, four bass columns
and six treble columns of 9 notes, less 5 at the treble end to give
85 overall, so 9 notes apart.  An arpeggio of this nature must be
virtually unknown in music and even if it weren't, and to avoid chance
occurrences of notes in that same pattern, Kastner only had to adjust
the note timing on the roll suitably by altering the tempo slightly.
(My earlier recollection of five plus five columns of eight notes is
evidently wrong.)

> The next question is how the accents were achieved.  The player stack
> was just an ordinary stack (Autopiano), or so it appears.  That means
> that the "individual accent" is going to be expressed across the entire
> stack at a time and then "bang-bang-bang-bang-bang" down the column of
> unwanted holes as well, even though it was intended for a specific note
> or chord (it would be interesting to hear what would happen during
> rhythmic arpeggios and runs up and down the keys, too).
>
> Possibly the stack was "split."  But the best you can get with a split
> stack is still no more ideal than any other reproducer can manage, and
> the true dynamics of the accents are impossible to achieve.

The stack was not ordinary at all.  It had two suction manifolds.
In an ordinary double-valve stack, the secondary valve operates as
a "changeover" between atmosphere and main (treble or bass) stack
suction.  In the Kastonome, the theme pouch (one per note actuated by

the additional ports in the tracker-bar) operated a second "changeover"
valve inserted in the suction supply to the usual secondary valve,
switching the supply from subdued main suction (lowered by operating
the "Kastonome" switch) to "theme" main suction.  Thus a theme pouch
operating by itself did not actuate the pneumatic.

With normal rolls, the "theme" suction manifold, pouches and valves
were never used.  The system thus provided instantaneous theme power on
any note without any need to delay or advance it as with Aeolians
"Themodist" and similar systems.

I have three or four Kastonome rolls here somewhere, survivors of a box
of 30 obtained in a job lot in the 1960s and swapped for plain rolls
with a collector who had a defunct Kastonome-fitted Grotrian-Steinweg
upright.  He never repaired it and when it was sold it went to a re-
storer who simply converted it to a plain player.  The rolls do turn
up in England quite a bit but they're mostly unexciting Edwardian pops
-- "The Arcadians", Poet & Peasant" -- that kind of thing.

> It is an amazing system, for sure.  This is a real collector's item,
> and I congratulate Inigo for finding it.  I hope that he can get it
> fully operational, once again, but I wouldn't want it for my first
> player piano.

I'd put up with quite a lot to get a Grotrian !

Dan Wilson
UK
Sat, 4 Dec 1999 19:27 +0000 (GMT)

The 'Kastonome' was an expression mechanism used on some player pianos
manufactured by the firm of Kastner, London, England.  It operates
quite differently from the Themodist system.  Instead, of only two
theme ports, as in the Themodist, the 'Kastonome' mechanism actually
employed 90 tracker bar holes, 85 of which were used for individually
accenting the 85 playing notes.  The remaining 5 were not used.  By
using one tracker bar theme hole for each playing note it was possible
to accent any number of playing notes simultaneously as playing the
unaccented notes.  When a theme hole is uncovered by the music role,
the accenting mechanism alters the suction level in the secondary valve
of the corresponding playing note.  This is achieved by a bank of
85 pouches connected to the theme holes.  A full description of the
mechanism is given in the book 'Pianola' by A Orde-Hume.

The 90 'Kastonome' tracker bar holes of ordinary size are located in
10 columns of 9 positioned on either side of the playing notes.  The
system utilizes the very low probability of a theme perforation
coinciding with the start of a playing note to prevent accidental
accenting.  It will be evident that most of the theme holes will be
vertically offset from the notes that they are associated with.

Four columns on the left (bass) side of the tracker bar, and six on the
right hand (treble) side.  The location of the bass theme holes is why
there is no sustain pedal perforations on 'Kastonome' music rolls.  The
first three columns of the treble theme are located where the top three
notes of an 88 note roll are located, and the middle row of each column
is located at the same level as the playing notes.  Interestingly, of
the three rolls that I managed to acquire for a Kastner 'Kastonome'
only one uses the bass theme.

For the record, the 'Kastonome' holes in the tracker bar are connected
to the playing notes in the following pattern:

 1  10  19  28                45  54  63  72  o   o
 2  11  20  29                44  53  62  71  o   o
 3  12  21  30                43  52  61  70  79  o
 4  13  22  31                42  51  60  69  78  85
 5  14  23  32  x .........y  41  50  59  68  77  84
 6  15  24  33                40  49  58  67  76  83
 7  16  25  34                39  48  57  66  75  82
 8  17  26  35                38  47  56  65  74  81
 9  18  27  36                37  46  55  64  73  80

x is playing note 1, y is playing note 85 and theme holes 41, 50, 59
are located at playing notes 86, 87 and 88.  Those shown as o above are
not used.  All tracker bar holes are of ordinary size and spacing.

While on this subject, a problem.  The 85 tubes for the mechanism
originate from a unit mounted behind the spool box.  This is located
approximately half an inch above, and hanging over, the back of the
hammer heads.  These need to be connected to the main stack immediately
above the playing pneumatics.  This location is underneath the back of
the hammer heads.  The original tubing was lead and curved behind the
hammers.  Even then I believe that central piano hammers rested against
it.  Due to the advanced state of disintegration of the lead,
presumably hastened by constantly being hit, it was replaced by
neoprene.  Unfortunately, the hammers tended to get caught in this, and
so it had to be removed (and not having any Kastonome rolls it was not
missed).  Would anyone like to suggest a possible solution of how to
run this tubing?  It would be nice to have the mechanism working.

I hope that the above is of interest.  The 'Kastonome' system has been
mentioned before in MMD, and the articles are indexed in the archive.

Douglas Bush
UK
Mon, 20 Sep 1999 21:47:17 +0100