Reading the Pinned Music of an Organ Barrel
by Craig Smith (050306 MMDigest)

smithc01.jpg (21 kb)
Figure 1. Transcribing the music of a 27-key Gavioli reed barrel organ.
The pinned cylinder is 37 cm (14.5") long and 15 cm (6") in diameter; the wood base of the
machine is 145 cm (57") long.  The multi-meter is used during setup to check the switch contacts.

Many months ago, I mentioned that I (with the help of Bob Cole) had built a reader for organ barrels [010303 MMDigest, "Connecting MIDI Wireline to Laptop Computer"].  I have used it several times to convert the pinning on the barrel to MIDI files.  I can use the resulting file for several things:

  - to play the music on a barrel if the rest of the organ is not working or has not yet been restored
  - to help in tuning if there is a scale written on the key frame
  - to try different tunings and pipe arrangements using the computer rather than handling the actual pipes.
  - to hear an orphan barrel
  - to help in restoring the key frame if severely damaged so the physical parameters are not certain
  - to verify assumptions about the pipe design/tuning/arrangement

It takes a about an hour to set it up with a new barrel.  There are actually several clues right on the barrel itself, like the spacing of the tune selection grooves.  The 'key frame' on the reader is adjustable up and down.  The ends can be tipped to get it level and adjusted to the right angle relative to the contact point on the barrel.  It can be moved left and right to align it with the first row of pins and then moved in small increments to select each tune.  Individual keys can be set to the correct level and the contact point is also adjustable. I have about 60 keys that I can slide onto the key bar but I only keep about 40 on there at a time.  I can position them on the bar at any spacing to match the barrel being read.

A large dividing head is used to support the barrel and to turn it at 1/2 revolution per minute by means of a toothed belt-drive from a Bodine motor.  Each key has a normally closed contact to a common ground, the switch is opened when the key is lifted by a pin on the cylinder.  These signals are directed to the input lines of an Octet Parallel-to-Serial (MIDI) converter that detects an opening contact and then sends the MIDI code into a laptop computer (via Cakewalk).  Then the program is used to assign each MIDI line to the 'correct' note so the music can be viewed and played.

smithc02.jpg (39 kb)
Figure 2. Detail of barrel for 27-key Gavioli reed organ.
The pins are 0.5 mm width, the tune tracks of the 3 tunes are separated axially about
4.5 mm (0.177"), and the reader keys are spaced apart 13.44 mm (0.529").

Each reader key can be positioned individually.   Also, each key can be adjusted for vertical position and for when the contact opens if the key is lifted by a pin on the barrel. And there are screw adjustments for left/right and up/down of the whole key frame.  I know it's adjusted when the scan produces real music instead of noise, and believe me, it doesn't take much error to make noise.  Since the organs are usually non-chromatic, one note off causes all the rest to shift odd amounts and the result is a real mess.  It makes the job much easier when I can recognize a tune from the tempo or by shutting off some of the notes.

The adjustment procedure depends on what information I have about the organ.  If the original keyframe is in good condition, I use that to set the key spacing.  I use the dimensions of the grooves on the barrel shaft to set the distance between tunes on the barrel.   Then I can either play the notes directly into a MIDI player and listen to the tune or record them into the computer and work with the MIDI file in Cakewalk.  The notes can be assigned to the keys during setup or I can assign the notes in sequence and assign the notes later in the computer.  If  I'm sure of the spacing, I can confirm the tuning of existing pipes or figure out what they should be.  If I know the notes, I can listen to the music and make sure that the keys are adjusted right.

The hardest situation is when I don't know either one and have to start guessing based on experience and whatever clues I can find.  That Gavioli barrel is a good example.  Someone tried to repair the organ and made a new key frame, but the distance between the keys was wrong when I tried to play the barrel -- the music was just a lot of noise.  So, were the notes (tuning) wrong or was the spacing wrong or was the barrel from the wrong organ?  I checked the reed tuning against the notes stamped on the reeds, against the notes the reeds play and against the tuning of two other Gavioli reed organs.  After transposition, they all matched.  The barrel was rather unusual with 26 notes plus bells and the bell track on the barrel looked right.  I also found Gavioli written on the barrel.

So the spacing was the only thing left to consider.  It was off by 0.010" so the last note on the barrel was from tune #3 when the first note was playing tune #1.

It seems like it should be an easy job -- just move the notes around until the tune sounds right -- but let me tell you, it's a lot harder than it looks and the fewer things you know about the organ the harder it is.  Even with clues like the order of pipes and some sizes, it can be a daunting task.

A couple of years ago I tried to decode a barrel from a small (11 note, I think) serinette for a fellow in England.  We spent hours on it and never did get it right.  On the other hand, I did get the right notes for a 20-key Beloudy organ (1800, from England) that I still haven't restored, but I know how great it will sound when I do!

Craig Smith
Upstate New York, USA
06 Mar 2005 20:48:33 -0500
23 Mar 2005 19:03:21 -0500

rev. 23 March 2005