Roll transcription and copying has cropped up a few times in the last
week or so in various forms, and was the subject of an active thread in
the MMD in December 96 - then it concentrated on 'hole for hole' copying.
It's an area I'm interested in, so here's my contribution to the debate.

As has so rightly been said, our player and reproducing pianos are roll
driven.  Without rolls the instruments will be dead -- and those rolls
are showing signs of self-destruction.  A roll transcription project must
therefore aim to preserve and replicate the rolls themselves, rather than
just on converting them to more modern formats.

To recap the old thread: Rolls for players were always 'synchronous' --
cut in a punch-advance-punch manner.  New rolls to date have pretty well
all been non-synchronous copies -- in other words, not copies of the
original roll itself, but sort-of copies of the performance with an error
of +/- one advance of the copying machine at best.  I think we all now
agree that they really aren't quite good enough musically, and are of no
use whatsoever as archival material.  The archival aspect should be the
target for a transcription project.

We have been assured that it's possible to produce hole-for-hole copies
of rolls, although in practice they seem to be very few and far between.
From reading the archives, it's clear the MMD has the collective resource
to convert the theory into general practice by pulling all the skills
together.

We need roll scanning/reading systems designed to identify individual
perforations and record the results in synchronous form (such as the file
format offered by Wayne Stahnke last November, for instance).  The file
must be portable across computer systems, partly to allow the maximum
number of people access to the data, and partly to avoid obsolescence.

Having a common file format enables the use of any roll reader or
perforator.  Obviously a synchronous master could be converted to MIDI,
but as MIDI is sequential it isn't suitable as the archival medium.  All
this has been discussed before, so I would like to know from those who
have been working on such ideas how they have been progressing over the
last 6 months.  Perhaps a 'recap' thread in the MMD would be of interest
to all?

If producing a synchronous master can be made as easy and as cheap as
producing a non-synchronous one, and the technology and ideas are freely
available, perhaps we could persuade all the present recutters to adopt
the technique.  Cutting a synchronous roll costs precisely the same as
cutting a non-synchronous one, so there should be no objections on those
grounds.

Present ordinary copying quality is very good, but also is obviously not
as good as it could be.  Technology moves on, and clearly there must be
some reason why present recutters do not make synchronous copies as a
matter of course, so that's where we should concentrate.  Encouraging
them to do better doesn't make us ungrateful -- indeed, assuming they
will _want_ to do better is a compliment!

After cracking the software and hardware problem, my proposal for a roll
transcription project would be as follows.  It would be a coordinated and
systematic effort to produce hole-by-hole computer masters of as many
rolls as possible, from every type of roll format.  I totally agree with
Bob Billings (MMD 970813) that we should forget about protecting and
holding tight to the information.  I would go further: both the file
format and the roll data should be non-proprietary and placed in the
public domain with the explicit aim of having them distributed as widely
as possible, using the capabilities of computers and the Internet for
what they're best at.  The wider the information is spread, the more
likely it is to survive, and survival is the primary aim.

How about a Web site for downloading roll data -- simply enter the make
and serial number and get the roll?  Magic!  Anyone could simply obtain
the required file and convert it into the form of their choice.  The
synchronous data could be used to perforate a new roll accurate
hole-for-hole, or be converted into MIDI, used to drive a player
synchronously via electronic valves, or anything else you can think of.
Making the data freely available is much more likely to encourage
perforator owners to adopt the new hole- by-hole copy standard, which
would greatly benefit all of us.

By coordination, duplicated efforts could be avoided.  If a database was
maintained of rolls available, reading resource could be targeted towards
un-copied rolls, so building up a complete catalogue.

This is a big area and I've gone on too long already although there's
lots more to say.  Let's hear from all the MMD contributors working in
this area -- it could be fun!

Julian Dyer