Julian Dyer wrote:

> My original concern after just hearing the CD was that the
> Boesendorfer, with its harder tone than older instruments,
> was rather over-accenting the music.

The Boesendorfer does have a more sinewy sound than most Steinways, but
Horowitz's specially maintained Steinway has a similar type of sound.

Transcription of 78s always seems to soften the sound.  78-rpm records
need a turntable with very good mounting to give their full dynamic and
avoid rumble.  Unfortunately the only Rachmaninoff 78s I have (as 78s)
are an incomplete set of the 2nd Concerto.  I would like to get hold of
some solo examples to compare with the transcriptions.

> The most questionable part of the CD is the speeds, which have
> often been changed quite a bit from the roll markings.  Given the
> compromises in performance when making a 78 record, it's questionable
> whether the 78 should form the definitive speed for the piece: the
> roll may well be a better representation, although roll companies
> could be pretty cavalier with the markings!  Again, it's an artistic
> judgement for the CD producer to make.

This question has been raised with other composers: was the performance
speeded up to fit a 78 side?  This seems very unlikely with a composer
of Rachmaninoff's integrity.  The main factor which would affect per-
formance speed, once the basic tempo is established, is the acoustic:
a large hall or church with long reverberation needs a slower perform-
ance than a studio or domestic living room.

Generally, tempos in the 1920s were faster than now.  Elgar's tempos
were fast:

  "Sir Edward has a drastic way of hacking at his music.  All sorts
   of things which other conductors carefully foster, he seems to leave
   to take their chance.  He hacks his way through in a fashion both
   nervous and decisive.  At the end we realise that detail and
   rhetorical niceties have been put in their place, and that the
   essential tale has been vividly told." (Musical Times 1926)

This is typical of composers playing their own music.  The difference
with Rachmaninoff was that he could play the details accurately at a
fast speed.

There is evidence that Elgar and Holst in particular did _not_ speed up
to fit sides.

The Horowitz performance of the Prelude Op 23 No 5 (on "Horowitz at the
Met") is a few seconds longer than Rachmaninoff's and there is a little
bit more of the trees and less of the forest, as you would expect.  The
Horowitz "Bumblebee" from 1932 is exactly the same timing.  As Horowitz
knew Rachmaninoff well, and both were Russian, he had a good under-
standing of how the music should be played.  The same strong accents
are there.

My guess would be that the engineers [at Ampico] set the suggested
roll speeds by trying out the rolls on a few pianos to see what the
mechanisms could handle (and perhaps what speed they liked best).

Regards

Don Cox
doncox@enterprise.net