Personal Comments on the T-100 Welte Recording Technology

Dr. Peetz, Craig Brougher and I spent a lot of time via e-mail trying
to unravel the Welte recording technology.  It was a delightful and
educational experience.  The basic problem was proving to ourselves
that the late Ben Hall's description so long ago contained some real
errors in his understanding of just how this system worked.
Unfortunately, people have just copied his description word-for-word
in their later publications.  [See footnote below.]

We had to first strictly use only electrical technology of 1903, and
not introduce more errors by going with later devices that were just
not available to the Welte firm.  Our conclusions were as follows:

1. The single mercury trough was probably wrong.  Splashing would
certainly be introduced with a fast piece of music.  Also having the
carbon rods just suspended by a spring would result in total loss of
any containment of the rod during movement.  Surface tension would
cause them to fly all over the place.  Not a likely thing at all to
have happened.  They had to be guided.

We were convinced that the mercury was either in small glass vials or,
if the trough was indeed used, transverse baffles had to have been used
between each carbon rod, plus some type of guide rail for the rods.
Most likely, the glass vials would accomplish both needs of mercury
containment and rod guidance.  A simple and most adequate way to
contain them.  At any rate, each carbon rod had to have been isolated
from adjoining rods and guided.

2. As Hall stated in his piece on the Welte system, the carbon rod
could not have been suspended on springs.  The bad dynamics introduced
by a suspension spring bouncing all over the place would have
completely upset the need to precisely record each note's dynamics.

Dr. Peetz correctly concluded that there was a spring-loaded push-rod
between the key and the carbon rod, to keep the rod assembly in contact
with the key under all conditions, and why it was needed.  What Hall
may have misunderstood is that there had to be a very fine coil of wire
connecting the carbon rod to the voltage source, and the compression
spring-electrical connection most likely was the one that Hall thought
was a suspension spring.

3. Mr. Simonton showed me not only some of the carbon rods but also
a short piece of the original recording paper roll from the master
recorder.  It was easily seen that fast notes showed a very fast rise
time in the ink trace at the beginning of that note, and a soft note
had a longer rise time.  The sharp rim of the soft rubber wheels did
this job, depending on how hard and fast it was pushed into contact
with the roll paper.  This was what the roll editor used to determine
the note dynamics.

4. What was never further worked on to any conclusion was Hall's
statement that the master roll could then be instantly played back 
for the musician, because the ink was conductive and some form of
electrical reader was used, that operated a special Vorsetzer for
playback.

I, for one, do not believe this at all.  It means conductive ink, and
passing any large amount of current through the ink would set the paper
on fire, if you were using linear solenoids.  You are talking about
a lot of current.  Not very practical, when considering the need for
varying pressure for expression.

Then, providing the varying dynamics in the special electrical
Vorsetzer would have been an engineering nightmare.  Ten organ magnets
for ten steps for every single note? Average the accompaniment and
melody sides first?  Not at all believable.  No, while possible to
actually do today, it was just not practical then, nor even really
needed.  The varying ink lines were read visually, no electrical
interface was used.

Plus, how would one precisely read the varying width of the note's
beginning electrically?  Hundreds of contacts across the line?  Not
likely.  Photocells were not usable in 1904 because all they then had
were selenium cells and they have a slow response.  The contact systems
to read the note markings that we envisioned in our deliberations,
using only known electrical means in 1904, were so involved that one
has to conclude it just was not done.

I have read that Edwin Welte himself and one lady were highly skilled
at reading these note markings and correctly punching in the right
expression codings.

The ink on Simonton's roll piece did _not_ look anything like it
contained added graphite or silver powder, or anything else except
plain black ink.  Of course, it was some fifty-five years old then.
Personally, I am convinced that the roll was edited, the expressions
put in and punched first, _then_ the artist came back to the Welte
studio and reviewed his performance.  Needed corrections were made
and the production roll master was made from that.

Whatever, the Welte T-100 system was an astonishing achievement,
quite equal to the 1915 Skinner Orchestrator organ roll player in it's
sophistication.  Both a tour de force of brilliant thinking and high
level engineering skill.

After owning two Duo-Art Steinway pianos and listening to many Ampico
systems, the original Welte system using a Vorsetzer and playing a 9'
Steinway grand was the most thrilling musical experience I ever had,
and I was laying on my back on the floor under the piano, too.  It
beats the others hands down.  Being present in the studio during the
Simonton recording sessions, and hearing that great Steinway and the
Vorsetzer doing what they do best, was the thrill of a lifetime that
I never will forget.

Jim Crank

 [ The description by Ben Hall appeared originally in the booklet
 [ that accompanied the LP album, "Welte Legacy of Piano Treasures",
 [ produced by Walter S. Heebner under license from Richard C. Simonton.
 [ It is reproduced on page 327 in Bowers' "Encyclopedia of Automatic
 [ Musical Instruments".
 [
 [ A similar description was published many years earlier, written by
 [ Albert Goldberg, music critic of the Los Angeles Times (from 1947)
 [ and author of the newspaper column, "The Sounding Board".
 [
 [ His article, entitled "In Ghostly Array", a review of the 1950 LP
 [ of Welte-Mignon music, was published in the L.A. Times on Sunday,
 [ April 9, 1950.  Goldberg didn't credit any other author so the
 [ technical description probably was provided by Richard C. Simonton.
 [
 [ Gerhard Dangel of Augustinermuseum kindly provided the images of
 [ Goldberg's article that are presented at
 [ http://mmd.foxtail.com/Pictures/Welte/lp1950review.html
 [
 [ -- Robbie