I'm late responding to a question about recording music boxes, but
I hope the following may be of assistance.

The co-founders of our Yesteryear Museum in New Jersey were George and
Madeleine Brown, of Chatham and later Morristown, New Jersey.  They did
wonderful radio programs and public appearances with music boxes and
allied material for many years, starting about the time of World War
II.

I met them when they gave a junior high school assembly program around
1948 or '49.  I was in charge of stage arrangements.  Well, there you
go again blabbermouth, giving away your age!  We retained our
friendship for many years, culminating in the collaboration on
Yesteryear during the 1970s.

The Browns' collection was featured on several fine LPs put out on the
Cook label (the original stereo LP company) and later on Caedmon,
otherwise better known for spoken-word anthologies.

I suspect that the Browns added their own expertise to suggestions made
by Ruth Bornand, the originator of remarkably good music box phonograph
recordings in the 78-rpm era.  The early ones were later transcribed to
LPs, then to audio cassettes, where they joined newer siblings, and I
assume later to CDs.

One of the things they discovered in recording a music box was to place
the instrument on a sturdy box with the microphone _underneath_ the
music box.  This helped reduce the interference of spring chugging and
other motor and governor noises, and provided an additional sound
chamber with the bottom of the music box above the mike, which was set
on a cushion below.

To varying extents, musical box movements have a connection (usually a
wooden bar) joining the mechanism bedplate to the bottom panel of the
case.  This "telegraphed" the sound from the combs to the bottom, which
acted like a sounding board, as on a piano.  Using the microphone-in-a-box
technique also helped muffle or eliminate extraneous outside noise.

One tends to ignore such external noises while setting up and
recording, until one hears them on playback.  Listen to some of the
Rita Ford LPs -- they were made in her retail store located on a busy
New York street, with truck noises, horns, and the like.  Realism!
Annoying!  Unprofessional!  Shame on Columbia Records, Incorporated!

Unfortunately, these are the discs which attained the widest circulation,
because of the size and distribution of Columbia Records.  I suspect
this turned off a lot of people who might have become enamored of
mechanical music as are we, had they been exposed to the Bornand, Cook
or Caedmon releases.

The second discovery for recording involved disc music boxes, where the
rotating disc will attempt to hold its "dished" shape.  This is where
the part of the disc moving over the starwheels is pushed down to
engage the teeth of the starwheels.  At the same time, the rest of the
disc is pushed up by little casters, away from the bolts holding the
combs in place, and other parts of the mechanism.  Most of its life,
the disc sits unused in this vise, while that "warp" becomes ingrained.

I kept my disc boxes with the hold-down bar raised to alleviate this
condition.  Without this caution, when the mechanism is turned on and
the disc turns, the effect is to move that depressed line around the
disc, which it resists.  As this movement takes place, the disc will
snap into the new position, much like a steel barrel-head will snap
when pushed in and again when released, like a drumhead.

Clockwork springs will likewise tend to stay in one position, and will
store up energy and then snap into the new position required by their
unwinding action.  Not only can this produce an annoying noise, but a
severe vibration in the mechanism.

This vibration may not be too important in a music box; it is most
annoying in spring-drive phonographs.  This problem with the Edison
Long Playing Record players drove project supervisor Theodore Edison to
distraction.  It would knock the sensitive stylus from one extremely
fine groove to another.  This is one of the reasons why the Edison LPs
were withdrawn.  Although capable of recordings far longer than the
earlier RCA Victor "Program Transcriptions" and the subsequent Columbia
LP, they were not practical.

Back to recording the disc music box.  The solution to the disc-snapping
is to get a heavy pair of gloves.  Better yet, the kind of car wash
mittens with thick shaggy surfaces.  Put them on your hands during the
recording session, and "walk" along the disc as it turns, one hand in
the front, the other to the rear.  Be careful not to bend the disc
down, so it won't catch on the protruding mechanism parts mentioned
above.  Experimentation is the key to learn how much pressure is
correct.

Another trick is to _not_ wind the box fully, nor let it run down too
much.  Experiment to find what is sufficient winding to play the length
of the tune one wishes to record, and wind to that point only, for each
selection to be recorded.  Perhaps the reader has caught on that as in
so many undertakings, preparation is the key to success.  Much of it
needs to be done before the recorder is ever turned on.  These
suggestions will make for superior recordings.

Then, of course, there is the whole story of using the right kind of
microphone, amplifier and recorder.  Here again, experimentation is the
key.  Don't utilize cheap equipment; here as in most things, one gets
what one pays for!  Do _not_ over-modulate in recording.  If anything,
err toward lower recording volume.  Playback volume can always be
adjusted in making the master production record.

I'm a radio-program fan; so many shows recorded on early tape-recorders
are so over-modulated as to be useless.  Unfortunately, they are
usually the only such recordings which exist.  Many recordings of
mechanical musical machines suffer from the same shortcoming.
Motto: "Less is more"!

I hope all this is of help.  Should anyone have questions or other
suggestions, I'd love to hear such.

Lee Munsick
Virginia USA