My name is Spencer Chase and I am new to the Automatic Music Group
and to the internet.  I have received most of the last years digests
from a friend and have some idea of the areas of discussion.  I would
like to present some of my interests in the hope that others share some
of them.

   So far my experience with automatic musical instruments is limited
to Duo-Art reproducing pianos.  I have been collecting and working on
them for about five years.  My goal is to eventually coax every nuance
the artists and editors put into the rolls out again.  My feeling is
that many of the recordings are excellent although most of the pianos
today do them little justice.  I currently have three pianos that vary
in their condition, and degree and quality of restoration.  I hope to
eventually restore them to "factory" or better condition.  I am also
very slowly building a Duo-Art vorsetser based on a late model fan
regulator and an Ampico stack.  I have received varying opinions as to
whether there is any chance of this producing musical results since the
two systems work on somewhat different principles.

   Two of my pianos are Steinways with the infamous cross valves.
After polling everyone I could find who was familiar with Duo-Arts I
decided to rebuild the valves of my first Steinway using the original
parts.  Although my piano performs better in most respects than most
that I have heard, I am not happy with the results.  I searched hard
and long for a leather that I felt would be the best from the available
choices.  I found some kangaroo that I thought had the right properties
but it was too thin and I did not want to take chances laminating it to
something else.  I used a firm and smooth Italian handbag leather that
I thought was firm enough to not take an impression from the seat.   I
tested it in mocked up valves and it seemed fine.  In the piano it is
another story.  The leather never seated in and I get too much top seat
leakage at low playing levels to reproduce proper dynamics of varying
sized chords.   It is my understanding that the roll editors expected
about a .25 wci drop in stack vacuum per note played at or near zero
level and whatever that translated into at other levels as determined
by the properties of the leather and everything else in the highly
interdependent non feedback type regulator (there has to be a better
term for the Duo-Art control) system of the Duo-Art system.  My valves
drop an average of .4 to .6 with each  valve on at a 4.8 wci zero
setting (I can't set it any lower without losing toughly coded notes,
which is a problem in a small room.)  This causes an annoying wimpiness
in some soft passages.  It can be improved with a couple of glasses of
wine, but I refuse to allow my piano to drive me to drink.  I hope to
get better results with the next Steinway.  I would like to get
opinions on types of leather to try and results that have been obtained
with round retrofit valve plates.  This brings me to my next topic.

   KANGAROO LEATHER LOOKS GREAT AND IT IS AVAILABLE.  There seems to be
a lot of discussion about Kangaroo leather lately but everyone seems to
be wondering where to get it.  Schaff Piano Supply has it.  I have
samples and the stuff looks great.  They also assured me (today) that
they have a current stock and steady supply.  They have a pouch leather
that measures .008 to .010.  It is smooth dense and supple.  Is nothing
like the fluffy porous sheep stuff that you could easily poke you
finger through.  It does not have pinholes.  The color is lovely too,
although only the moths would know.  They also have a .045 thick
leather suitable for valve faces.  It is strong and supple and has the
smoothest (without being squishy) nap I have ever seen on leather.  At
60x binocular magnification it looks like a golf course, not like a
shag rug as do  most other leathers.  The only problem is that Schaff
does not sell to the public.  THEY ARE STRICTLY A WHOLESALE SUPPLIER TO
THE TRADE.  They do not want to be bothered by hobbyists buying
retail.  They do not have a retail outlet of any kind.  They only way
to buy from them is through a piano technician or a retail piano store
owner who is willing to place an order for you if you do not qualify
yourself.  They also have excellent fabric covered pole hose and some
very fine pneumatic cloth in a few weights.  These three hard to find
items may make it worthwhile to befriend a local piano technician.  The
LEATHER SUPPLY HOUSE 219 234-8990 also purports to have KANGAROO pouch
and valve leather.  They sent me a card saying so, but I have not seen
samples.  They will sell to anybody.

   My next major area of interest is the preservation and duplication
of original piano rolls.  I would like to build a scanner and punch
eventually but this may have to wait for my next lifetime.  I have read
with interest the postings regarding these topics and would like to see
the discussion continue and would also like to see a collaborative
effort to develop a system that can be widely used to preserve this
heritage.  I'm sure the efforts of the members of this group could
produce an inexpensive and accurate recording system that could be used
to this end.  I remember reading about ideas for using readily
available hardware for optical scanning.  It seemed that line scanners
were too narrow and that ccd frame devices were too slow to use for a
single scan line unless the vertical element was suppressed which
seemed like a difficult reverse engineering job.   Also frame grabbing
and advancing sounds like a mechanical nightmare.  A thought occurred to
me that I discussed with my father, who is an electronic
instrumentation engineer.  We worked through the numbers and it looks
like real time scanning should be no problem with cheaply available
ordinary stuff.  The idea is to suppress the vertical element
optically.  Illuminating the roll with monochrome light and masking a
narrow window (eg.  narrow enough to detect chain bridges) would allow
a cheap cylindrical lens to spread the scan line over the whole ccd
frame.  This would turn a row of dots into a series of vertical bars.
Therefore any scan line from the ccd would be the same and the scan
rate would be the line rate instead of the frame rate.  This speed
would allow about 4x scanning (though I doubt the paper could take it)
or lots of redundant information for checking errors or bad pixels.  I
tried a crude experiment and even a piece of Pyrex rod seems to do the
basic optical part.  If there were optical distortions I guess these
could be adjusted in software, but I don't think it would be
necessary.  Has anyone tried this or thought of it ?

   If there is anyone still awake at this point, I promise to be more
concise the next time.  I have been trying to get onto this discussion
group for some time and I couldn't control myself.  I'm sure there are
other things I forgot to mention or ask.