This is some thoughts and questions about musical aspects of
reproducing player piano music.  Thomas Henden uploaded a bunch of
MP3 audio files to his web site a few days ago (Recordings Of My
Duo-Art Piano, 020227 MMDigest).  When I downloaded and listened
to them, some questions started to turn around in my head, which
I would like to try to write down.

Essentially, the main question is: Why is the quality of the
arrangements for reproducing piano so varying?  And quite markedly
so, in my opinion.

After an initial error of mine (I thought that the files marked "Weber"
were the inferior ones; I'm sorry that Mr. Henden had to read a harsh
e-mail from me, which lead him, among other things, to correct me on
this), I would just like to outline a few thoughts and interpose some
questions about the recordings marked "Weber" and, additionally, the
recording of "Ain't She Sweet", which is not marked "Weber".  The order
is alphabetical according to title.

 * "Ain't She Sweet"

A really good arrangement, in my opinion.  It is an almost
"Gershwinesque" quick-step arrangement.  It uses a player piano like
an orchestra, for what usually can only be accomplished by four hands
playing.  Is this arrangement for four hands?  I'm quite sure this is
so.  There is the stride in the lower notes, then melody, sometimes
counter-melody, and finally figurations above the melody -- at least
four "functional voices" altogether almost all of the time.  What makes
this arrangement so good is the "double function" of the functional
voices -- they not only "play their role", but also "grow out of each
other": the bass gets (counter-)melodic at times, the figurations flow
into the melody and vice versa etc.  Thus, each voice "makes more
sense" (literally!) than if it just would do its job.

 [ This is Pauline Alpert's splendid arrangement of "Ain't She Sweet",
 [ released in both 88-note and Duo-Art versions in USA and UK.  I have
 [ the UK Metrostyle issue (coupled with "Thinking of You") and it's
 [ lots of fun to pump!  -- Robbie

The interpretation also uses a reproducer perfectly, in my opinion,
first just to differentiate between the stride bass + accompaniment,
melody, and figurations, but at many places also for subtle effects.
Listen to the beginning of the verse (about at 0:54): When the second
note (the syncopation) is stressed just so little more than the notes
before and after it -- this makes me smile.  The stride is subdued
enough not to get into the way of the melody; the counter-melody and
figurations "flow into and out of the melody"; all the voices mesh
together really nice.  Besides, the basic note-lengths were given much
attention: subtle staccato and "drawing off" before syncopations,
legato to build tension etc., etc.

* "Fanfare for Organ"

Arrangement: Good.  Better even if you consider that arranging organ
pieces for piano is almost impossible.  Uses great tricks (low,
long-sustaining bass) to mimic the organ.  Of course, the 4-hand
arrangement (Did I hear this correctly?) allows for "fuller sound
spectrum".

Interpretation: Also very good.  I would have only one criticism,
concerning three(!) notes: After the middle part, at about 1:45, the
3-tone crescendo leading to the reprise sounds too "trivial".  There
would have to be a great ritardando there.  Nothing else that disturbs
my ear.  One could believe that the "intro" and the first bars are too
"trivial".  But after getting the "hang" of the arrangement - "for
organ!" -, one finds out that this is the correct way to go into this
piece.  One could also believe that the huge dynamic changes
contradict the "for organ", but exactly this makes an arrangement
excellent: if you mix the best of the original instrument and the
actual instrument.  A piano was called piano-forte when it was
invented, for an obvious reason; many interpretations do not take this
to heart (and think it's called "forte" today).

* "Going to Pieces"

Arrangement: Most of the time 2-hands, but occasionally, mixes in third
voice, so some editing was done.  Still, this is a far way off from
"Ain't She Sweet".  In general, only just two functional voices (stride
bass + harmonized melody).  (Is there a faulty note at the end of
second 0:36?  A sharp seems to be missing.)  When a counter-melody is
used later it is far away from the melody, so the two cannot "play
together" (one coming out of the other, for instance).  On the whole,
this is a "nice" arrangement, but in my opinion far away from really
good.

Also, the interpretation seems somewhat crude: It seems that the
expression is only assigned per phrase: 8 to 16 bars with the same
level.  Could it be that this roll was not made for reproducing piano
originally?  Has this been done?  It seems logical, a cheap way to
increase the repertoire for reproducers, maybe with some well-selling
titles.  (And my download suddenly breaks off in the middle; might be
my fault.)

* "Lion Tamer Rag"

This is a very standard ragtime arrangement, with some additional
editing (a 3rd voice at some places) and "octaving", but the
interpretation (expression) somehow transports "ragtime".  It's by far
not as complex as in "Ain't She Sweet", but a solid "straight"
stride + melody structure.  A question I would have is, why no legato
in ragtime?  It appears to me that someone transcribed the pure notes,
but did not add a sustain pedal (or a "sustain per note", where each
note is lengthend on the roll).  Could that have happened here?  Of
course, it might also be intentional, but it's so much against the old
rule of Scott Joplin on how to play ragtime ("Play ragtime melodious")
that something seems "missing".

* "Pickles and Peppers"

Also a very standard ragtime arrangement.  In my ears, it cries for
some counter-melody, at least in the repetitions, but that would
probably have made the arrangements much more expensive, money-wise.
However, there is a sudden single trill in the quiet part, at 1:46, and
a sort of banjo effect between 2:20 and 2:40, so some editing was done.
Why, then, not add more counter-melody, etc,?  (And - weird - the two
notes at the end!)

Also the interpretation is something which I do not really understand.
It gets more and more quiet, down to a very low level, e.g., at 3:16
it seems like someone wanted to state, "This is a reproducing piano".
Why do it like that?   Is it easier than doing a note-by-note
expression assignment?  Was this (also?) originally a standard player
piano roll, which was upgraded to reproducing piano?

* "Rats Rag"

This arrangement does use counter-melody heavily, although not very
sophisticated.  The "counter"-melody at many places is almost in
parallel to the melody proper.  This arrangement sounds like an
orchestrion arrangement re-written for piano; maybe the arranger wanted
to create "registers" ("stops") on the piano by moving the
counter-melody up so high.

The interpretation (use of expression) again seems somewhat crude,
taken from a standard player roll and retrofitted with "some"
expression?  However, the stride seems to be played with a somewhat
lower level than the melody, and the expression levels "make sense" --
they underscore what's going on (a sort of call and response).

Probably I have to stress after these opinions that it is not my
intention to write a negative critique about these rolls -- their
arrangers and players are probably long gone, and I certainly do not
want to criticize today's listeners for what they like or not.

What did make me curious (and, yes, during the first few times I
listened to them at least, "emotionally upset"), is simply the question
why these rolls are *so* different?  Does it have to do with how much
work was put into post-editing?  With the general attitude of the
producer of the rolls?  With the (good or bad) times for the
roll-producing industry?  Does it have to do with the time of the
recording?  "Ain't She Sweet" was composed in 1927, as far as I know;
the rags probably date from much earlier.  I would really like to know
more about these questions!

So much as a (maybe very personal) contribution to the musical aspects
of our hobby (or profession) ...

Finally, I would like to thank Thomas Henden very much for creating and
uploading these recordings.  For people like me, without a roll-playing
player piano at the moment, and with no reproducing piano, such
recordings are a rare occasion to get a glimpse of what some of you
can hear every day ... I envy you!

Regards

Harald M. Mueller
Grafing b. Muenchen - Germany
http://www.haraldmmueller.de

 [ You ask why the arrangement quality is variable.  I think it's no
 [ more variable than are phono recordings, with different artists
 [ and different arrangers and different musical philosophies.
 [ The piano roll industry issued music rolls by well-paid big-name
 [ artists as well as poorly-paid no-name artists.  -- Robbie