Philippe,  I have recently resumed transcribing pinned cylinders and,
yes, it is quite a tiresome problem, if you want to make a really
universal reader.  I have found a reliable solution for some problems,
such as making a general purpose cylinder transporter, for example; but
there are some aspects that are of really difficult solution.

One of the problems I was speculating recently is how to read the pins
without having any material contact with them.  I started to play with
the idea of using a modulated laser, and purchased a laser generator
and some optical stuff (lenses and special mirrors).

The procedure is not new; it has been used recently to make a map of
the surface of famous sculptures in museums as, for example, the
celebrated Davide of Michelangelo.  Of course, the tools they used
are very highly priced, but the principle can be realized homely at
affordable price.  Although I'm not sure it will lead to satisfying
results, it is quite a fun experience.

Mechanical reading of the pins is not enough if you need an accurate
transcription; often pins are bent and appear longer and displaced
after automatic reading, not to speak of missing pins or bridges, and
pins that break down as soon as you touch them.

In addition, in some cases the noteurs used to mark the holes for pins
with conventional signs that are of great importance for retrieving
complete musical information.  In those cases, an image of the cylinder
should be needed and compared with the pins reading.

For those reasons, I think that some hand work is unavoidable.  In
my opinion, the amount of work you want to spend in hand corrections
should be estimated according to the importance and the musical value
of the piece; in other words, hand corrections are worth the trouble
only if the cylinder really contains valuable musical information.

I hope I can report progress in the near future.

My best wishes of Merry Christmas and Happy New Year to all MMDers.

Leonardo Perretti
Italy