[ Ref. Wallace Venable in 230401 MMDigest ]

Regarding "AMI formats" and the Wurlitzer 125 band organ,
Mr. Venable is correct, not only is "the" AMI format not published,
there are several "AMI formats"!

The original AMI format was for the "Green Book" project to digitize
European fair organ and cafe organ music books. This format needed
3 channels to encode the notes and stop selections for organs of up
to 125 keys.

Then there is the AMI-2 format, which has some of the characteristics
of the "Green Book" format but encodes the W165 band organ music on 2
MIDI channels. I have seen this on a spreadsheet which includes slight
variations for the Stinson band organs.

There seem to be variations to accommodate Wurlitzer 150 and 125
formats, but I am not sure they can even be considered "AMI" formats.
I prefer to deal with roll recordings in "tracker bar order", that is,
the original tracker bar holes mapped to MIDI notes in the original
order on the roll.

Often the first MIDI note will not be note 0 or note 1, but will start
at MIDI note 36 (two octaves below middle C). These tracker bar ordered
files cannot usually be played in a MIDI editor because the original
scales were not chromatic. The advantage is that the original tracker bar
information is preserved, and then can be processed for any purpose.

Terry Smythe has performed a great service to the hobby by scanning a
huge number of rolls from a variety of sources. Very often he publishes
them in tracker bar order, if they will fit on the 128 notes of one MIDI
channel.

The Style 125 band organ files are the exception. The holes for Bass,
Accompaniment, Melody, Trumpet, and Percussion (channel 10) are split
among different channels so that they can be played from an editor or
MIDI file player. This would be easy to map back to tracker bar order
or any order for output to valves on a real W125 _except_ that Terry
did one more thing.

The General MIDI Specification for percussions on channel 10 does not
include a Snare Roll. The snare drum holes for W125 are indeed for a
reiterating "snare roll". Terry created multiple strikes of the MIDI
snare to match the extended 125 holes. Unfortunately this means that
the individual snare strikes have to be put back together to make the
original Wurlitzer holes.

I have done this for myself simply because I prefer to store roll scans
in the original tracker bar order. It also means that the files can be
easily mapped to whatever order might be needed to interface to a real
machine.

This, ultimately, is what Steve Muscato was talking about. His format
is one of these "sort of AMI" variations and is quite straight forward.
I have furnished him with some examples of TS files and some digitized
by the late Russ Doering.

I encourage anyone scanning or digitizing rolls to preserve all the hole
information on the original roll. If the information must be arranged
differently in the MIDI file, make sure the arrangement is clearly
documented. Later users will be very grateful.

Craig Williams