Hi, all!  Gordon Stelter asks, "Has anyone successfully transformed
scanned piano rolls to sheet music, and if so, who?"

Short answers: "Yes, and many people, including me."

Longer answer -- People have been writing out piano rolls [to sheet
music] for many, many years.  Originally, it was a rather painful
process.  In the most traditional method, employed by our Humble Editor
for example, one puts the roll on a foot-pumped player and plays a few
seconds-worth at a very slow paper speed and stops, and writes down on
manuscript paper what was just played.  I found from experience that
this is really hard on the main bellows of one's piano, though.

I used an even more painful method starting in the mid-1970s, when
I didn't have a player piano, wherein one unrolls part of the roll on
a table and measures the holes with a scale stick.  It really, really
helps in this case to have a recording of the roll being played for
reference.  As painful as it is, I found this method to be much more
satisfying than transcribing audio, as there is a definite correct
answer to be had, while in the case of transcribing old records, one
never can be sure if the transcription is really accurate.

But in the modern era, roughly the 21st century, things are a lot
easier.  Really, the explosion of roll scanners was started by
Richard Stibbons, of the UK.  His website is http://www.stibbons.com/ 
He developed a much simpler roll scanner, based on the Contact Image
Sensor (CIS), found in some kinds of flatbed scanners.

The CIS is essentially a one-dimensional camera, in principle very
similar to the way the tracker bar can be thought of as a low
resolution one-dimensional digital imaging device.  After Stibbons
generously published all of the information needed to construct
a simple, successful roll scanner, many, many people followed the
recipe with variations and fielded scanners that convert the piano
roll into a small MIDI file.

Converting the MIDI file into sheet music is a fairly straightforward
exercise, but a certain level of musicianship is necessary to produce
a readable, playable score.  It can be done with almost any music
notation program, in that most such programs (should I use the modern
term "app" instead of "program"?) have the ability to transcribe MIDI
files to some kind of notation directly.  But a lot of work is necessary
to align the file to a metric grid and edit it into something that is
usable, readable, and playable.

I have published a volume of transcriptions done this way of James P.
Johnson's rolls, and earlier completed a still-unpublished volume of
Johnson's rolls of his own compositions.  (That earlier volume was
finished in 2001, but publication has been on hold due to a thicket of
copyright and ownership issues.)  If anyone is interested in the later
volume, let me know and I'll publish a second edition if there is
sufficient interest.  I also have published a volume of transcriptions
of rolls by various players called "Piano Roll Sampler".

Nathan Bello has transcribed many dozens of rolls by the Southside
Chicago pianists, like James Blythe and Clarence "Jelly" Johnson,
and I believe you can purchase those from him.  His live recordings
are at  http://www.cdbaby.com/Artist/NathanBello 

Others have concentrated on transcribing rolls of compositions that
never were published in the form of sheet music back in the era, like
the many Charley Straight original rag rolls.  There is a list of
rags that fell into that category in John Hasse's book "Ragtime: Its
History, Composers, and Music" (1985) and I am happy to say that by
2015 I believe every piece on that list has been transcribed by someone
or other, though in many cases those transcriptions are hard to find.
One most often needs to ask the 'grapevine'.

I do believe that there is another way that people have transcribed
rolls, judging from close examination of the transcriptions.  That is,
to do a transcription from audio of a recording of the roll.  People
who do a lot of transcription from audio and are good at it are
naturally inclined to avoid a lot of the intricacies of dealing with
rolls by the other methods.  The one great advantage of this method is
that one does not need to have a copy of the roll, nor a player piano.

And very many rolls can be heard in YouTube videos of varying quality.
The great disadvantage of this method is the fact that it's hard to get
the details right, and it's impossible to know that you've gotten them
right.

Very best regards,

Bob Pinsker
San Diego, California
http://sandiegoragtime.com/