Re: Karl Petersen's "Discovery" Letters

Well, Karl sure did pull a good one on all of us.  Reading his post did
get me going, as his hypothesis had enough reasonable engineering facts
that maybe it could have described the Welte T-100 system.

I still dispute the often-stated "fact" that these inked master rolls
were ever used for any instant playback for the artist.  Just how that
supposed conducting ink-line could have been read electrically in those
days is a matter of unnecessary guessing.  It was not done.  There is
just no way to do this with great reliability and with consistency
using strictly 1902-1903 known technology.  Are you going to electri-
cally isolate each one of those tiny hair-wires in some brushes and
average the result while reading across the note line?  I seriously
doubt this idea existed even experimentally.  And why would you risk
any possible damage to the master roll?

Edwin Welte and Karl Bockisch were able to read the initial striking
force of a note by measuring the angle at the start of the note opti-
cally.  They had the calibration all worked out as to what angle
represented which expression-control holes to be used for the Vor-
setzer.  Then the master roll was punched for notes and expression
control holes.  Now, that roll could have been used in the playback
Vorsetzer, or--and I still think this is how it was done--it was
duplicated and _that_ roll was what was played back for the artist to
hear.  Then any corrections were made on that roll and not the master
roll.  Why risk the original that cost so much to make, what with
artist fees, production costs, and all that?  Then that corrected roll
was used to make duplicate rolls for the production punching machines
in the factory.  Both the inked master roll and the corrected roll were
then archived.

This master roll certainly could have been recorded at a faster speed
than the finished playback roll.  This was common practice with many
roll-mastering systems, as it gives one much better resolution to work
with for editing purposes.  The playback Vorsetzer certainly could
easily have been modified to run at a faster speed.  The many photos in
the Welte studio showing the artist listening to his roll show a
normal-sized Vorsetzer.  Putting some involved and complicated electri-
cal reading system into one would certainly make it much larger, as big
as a house!  And, if the reading contraption was remote, then the cable
to the Vorsetzer would be a big one, and not the simple small one you
can see in some of the photos, which looks like a normal power cord,
not some big, fat cable.

Karl's explanation of why that recording machine one sees in the back-
ground of a lot of pictures in the recording studio is so tall is prob-
ably quite right.  There is a need to let the ink dry before rolling it
up, and that certainly would explain why the machine was so unusually
tall: a drying cabinet.  It makes sense.

Dr. Peetz's concept of the key/carbon-rod mechanism makes great sense
and would work as Welte intended.

The Welte T-100 piano system did use sharp-edged, inked rubber wheels.
If the supposition is true that they were run by linear solenoids--as
I believe; the technology was known then--the inked line was constant
in width as long as the note was being held down.  Such solenoids are
linearly responsive to the increasing voltage, which the carbon-rod key
system did provide.  Common electromagnets with iron armatures do not
respond this way; they are more of a slam-bang action, fine for organ
rolls, which are either ON or OFF, but not for the velocity-sensitive
piano-roll recorder.

However, what I saw at Dick Simonton's home so long ago was this, to
the best of my recollection:

1) The carbon rods did not have any sharp tip, just flat.  If there was
something done to the rods before use, then these were just spares and
had not yet been shaped.

2) The original master roll was only a small length, maybe a foot or
so.  You _could_ see variations in the beginning of the notes.

3) Whether the lines were of varying width after the note was struck, I
just cannot recall, and now I don't think they were varying.  Once
down, the line only had to indicate how long that note was being held.
That beginning angle told all you needed to know about how hard the
note was being struck.

As several concert pianists have explained, they vary only the initial
strike force as they play.  Notes that are simply brushed enough to get
them to play are another matter; but it does not alter the fact that
the Welte system could see this.  The strike angle would differ, and
the line would tell how long the softly-played note was in action.
Reading strictly the hammer velocity is certainly the most accurate way
to determine note intensity; but nothing in those days would do this.
Today this is a different matter, and it can be done.

4) There was, to my best recollection, no way to tell whether the ink
was conducting ink or just plain black ink.

5) I seem to recall that Dick Simonton showed me some punched rolls
with all sorts of pasted-on corrections, which he described as "master
rolls."  However, these rolls did not have any inked lines.  The paper
was ruled, I believe.  These certainly could have been the second gen-
eration rolls.  They were definitely not the inked master rolls.

Anyhow, this has all been great fun.  But it does raise the question
whether a paper should not be written again on how the Welte T-100
recording system did work, strictly staying only with known 1902-1903
technology and using a good deal of engineering sense as to what would
be most reliable and repeatable.  I think we have all back-engineered
the Welte system enough now to take another crack at it.  Too many
false concepts have been put forward already.

Jim Crank