I've not heard many here in the UK who have anything but praise
for Wayne's CD of Rachmaninoff's Ampico rolls. These people aren't
uncritically praising it either, they are listening and comparing it
to real live performances for musicality, old 78s for authenticity,
and other recordings of the same Ampico rolls for comparison. Wayne
shouldn't be too disheartened, or even annoyed, by the one or two whose
minds and ears are closed to all but the noises from their own grinding
I hope my questioning of the roll speeds wasn't seen as a criticism --
it was a question, and has been answered in a fascinating manner by
several contributors. Now I wonder why the Ampico rolls have tempos
that are so open to question!
I was very interested by Wayne's research showing that Ampico rolls
pre-1925 were not recorded with in-built acceleration compensation.
Rex Lawson did some major research when this same assertion was made
regarding Duo-Art rolls in the Nimbus series. All the evidence,
from existing perforator original rolls, patent drawings, whatever,
suggested quite strongly that Duo-Art rolls were acceleration-compen-
sated from the start. It would seem extraordinary if the Ampico rolls
weren't, yet I'm sure Wayne's research is accurate. Does anyone have
good evidence to the contrary?
All this is interesting regarding 'authenticity' and the search for
it. Duo-Art tried some form of real-time notation of the performance
dynamics almost from the start of their recording program, and gradual-
ly refined the way in which these notations were interpreted to code
the dynamics on the rolls. They also had the in-built acceleration
compensation. Their product tried from the start to represent the
performance using reasonably objective techniques. Yet, on top of
knowing Ampico did not record actual dynamics until the late 20s (as
the roll editors interviewed by Nelson Barden emphatically stated), we
are now being told the note record had systematic inaccuracies as well!