Dr. Clarence Hickman devised a velocity trap timer to measure and
record the hammer velocity of each piano note as it was played.
A pair of wire contacts were fitted to each hammer of the piano and
connected to the marking recorder where spark coils generated enough
energy to melt a tiny hole in the waxed paper roll.  After the
recording session the paper was dipped in an ink bath, so ink would
enter the hole in the wax and color the paper where it was struck by
the sparks.

The distance between the two marks on the waxed paper was proportional
to the elapsed time for the hammer to travel between the two contact
pairs.  The music editors measured the distance on the waxed paper and
converted the data into Ampico expression codes.  Hickman varied the
contact distance from bass to treble so that the recorded data was
really related to the energy of the hammer.

Dr. Hickman wrote in 1929:

  "While this intensity range of 18 decibells may seem small, it should
be remembered that this range is for a single note.  When an artist is
playing fortissimo and using the sustaining pedal, it is possible to
build up an [sound] intensity perhaps twenty times as great, which would
add an intensity of about 13 more decibells to the range of the piano,
giving a total of about 31 decibells.

  "It has been shown that the velocity of a bass hammer is less than
the velocity of a treble hammer for the same level of loudness.  This
change in hammer velocity is not an abrupt one but takes place gradually
from one end of the piano to the other.  This is to be expected since
the hammers are all different in size.  They are moulded in a tapering
form and then later sawn apart.

  "In order to avoid the necessity of using a different measuring
scale for each hammer, the contact distances in the bass portion of the
recording piano are made smaller than those in the treble.  Each hammer
has a different contact distance.  The total variation from note #1 to
note #88 is about .030 of an inch.  This gives approximately .0003 inch
difference in contact distance for adjacent notes."

(Quoted from page 145, J.A.S.A, Vol. 1, No. 1, October 1929.)

Hickman didn't give the mounting details nor the actual contact
distance for the hammers, but it's assumed that the wire contacts made
connection after the jack letoff and just before the hammer struck the
strings, e.g., during the last 1/8-inch or less of the hammer travel.
Maybe there is mention of this in the Nelson Barden interviews.

Essentially the same method is employed in the Boesendorfer SE and
Yamaha Disklavier recording pianos, but instead of two wire contacts,
a tiny shutter interrupts the light beam of a photo detector (or a
pair of photodetectors) just before the hammer strikes the strings.
The shutter is attached to the hammer shank near the hammer.  In the
SE system the shutter height is 2.5 mm (0.1 inch); the Disklavier
shutter is probably of similar size.

Robbie Rhodes

 [ More Ampico data at http://mmd.foxtail.com/Ampico/  -- Robbie