If you can’t beat’em,
join’em! Unable to prevent Naxos’s (and
others’) wholly legitimate and hugely
successful transferring to CD of some
of the plums from their catalogue as
year by year they come out of copyright.
Having lost a lawsuit in the United
States, EMI is responding as it should
in a free market economy, by issuing
its own transfers at a price similar
to Naxos’s and trusting that the public
will feel that, economic considerations
being equal, the owner of the master
tape is likely to do the best job.
My review of the Naxos
transfer was posted
on the site only recently so, as
far as the performance is concerned,
I will refer readers to my comments
there. They will also find at the foot
of the review a considerable exchange
of opinions between the transfer engineer,
Mark Obert-Thorn, Robert E. Seletsky,
author of an article on Callas at EMI
as well as a note in the Naxos booklet
and myself.
Basically, the debate
centres around Obert-Thorn’s and Seletsky’s
claim that the latest EMI transfer,
in the Great Recordings of the Century
series, "is noticeably flat, an
error which, in addition to adding nearly
a minute and a half to the running time
of this relatively brief opera, also
affects the listener’s perception of
tempo and vocal timbres" (from
Obert-Thorn’s note).
Relying on my unaided
ear, I was puzzled that the Naxos recording
seemed scarcely sharper than my piano,
which I knew had slipped to 438 Hertz
if not lower. In view of the hornets’
nest this stirred up, something more
than my unaided ear seemed called for.
Two people, a singer friend and my piano
tuner, brought round their pitch measuring
equipment and spent some time patiently
comparing selected passages (mostly
from "Vissi d’arte" since
long-held notes are required for a reliable
reading).
Before coming to their
verdict, however, the actual timings
would seem at first to bear out Obert-Thorn’s
and Seletsky’s contentions: the EMI
transfer takes 109:34 compared with
Naxos’s 108:31, over a minute’s difference
if not quite the minute and a half claimed.
However, if we look at some individual
track timings, the Naxos is not always
the shorter. Take the last part of Act
Three:
|
Naxos |
EMI |
Senti, l’ora
è vicina
|
1:30 |
1:30 |
Amaro sol per te m’era il morire |
1:56 |
1:54 |
E non giungono |
2:26 |
2:26 |
Come è lunga l’attesa! |
2:15 |
2:15 |
Presto! Su, Mario! |
1:20 |
1:17 |
(In some cases the
variations may regard the decision over
where exactly to start the new track).
Whereas the opening
sequence of Act One tells a different
story:
|
Naxos |
EMI |
Ah! Finalmente!
|
2:02 |
2:07 |
E sempre lava |
2:13 |
2:15 |
Sante ampolle! Il suo ritratto! |
1:13 |
1:14 |
Dammi i colori … recondita armonia |
4:20 |
4:25 |
Gente là dentro |
1:10 |
1:10 |
Now my naked ear told
me that the EMI version was very fractionally
lower pitched at most of my selected
points of comparison; but it also had
more body and depth to the sound, which
might have influenced my perceptions.
So what did my technological aids reveal?
As far as Vissi d’arte
was concerned, the equipment noted no
difference between the two and – and
here is the interesting point – gave
a reading of 444 Hertz for both of them,
in spite of Obert-Thorn’s and Seletsky’s
claims that the pitch had been corrected
to 440. It also revealed that Callas
was about a Hertz sharp on most of her
high notes. Another recording made in
Italy at about the same time, under
Basile for Cetra, also gave a reading
of 444 for the orchestra but this time
the singer, Gigliola Frazzoni, was about
a Hertz sharp right through. Kiri Te
Kanawa’s 1981 version, made in London,
likewise gave a reading of 444 and seemed
the purest of intonation of them all;
her 1996 version, made in Lyons, gave
a pitch of 445. No wonder she made heavy
weather of it!
So, if we assume that
two pieces of equipment which give identical
readings must be proved right, the possibilities
are:
- That neither Naxos or EMI have transferred
the recording at the correct pitch,
despite the protestations of the former.
- That my CD player mysteriously raises
the pitch of everything it plays by
about 4 Hertz. These was common enough
with LP turntables and cassette players,
but we are told it is impossible with
a CD. Still, they told us at the start
that they were unbreakable and everlasting
…
- That the 444 pitch was that of the
performance itself, so both transfers
are about right and the whole 440
business is a red herring.
Point 2) is the easiest
to dismiss. We tested the pitch of my
own CD of Cyril Scott, recorded in Milan
on a Steinway which my piano tuner has
been tuning regularly for some years
and which he assures me is kept at 440.
The CD played at exactly 440! We also
tested a demo-CD made recently by one
of my singer-friends, in another studio
near Milan. It played very fractionally
below 440, nearer to 439. There is no
reason to suppose this did not represent
the reality. In other words, if a recording
plays at 444 on my CD-player, it means
the manufacturer has transferred it
at that pitch.
Point 3) is more difficult
to establish. My piano tuner was kind
enough to contact some of his older
colleagues, and obtained the confirmation
that Italy did respect the 1939 international
convention (held in London) which established
440 as the concert pitch, which was
previously around 432. 440 was laid
down again in 1955. So a recording made
at La Scala in 1953 should have
been at 440. On the other hand, Seletsky
tells us that recordings made in Italy
in the early 1950s often did have a
higher pitch, and it is well-known that
in Vienna and many German cities a higher
pitch was regularly used until a further
international convention in the 1970s
tried once again to fix 440, which is
what is used in Italy today. So maybe
there was an "unofficial"
tendency to tune sharp in those days,
but as yet I have failed to find anyone
whose memory goes back that far. There
is also the question of how much pitch
might vary during sessions as the instruments
warm up.
Does this matter? Well,
since the two transfers differ only
minimally in pitch, whatever Naxos claim
to the contrary, then this can hardly
be an issue in choosing which to buy.
If it were to be decided to jack both
of them down to 440, I should think
the difference would be quite drastic
and I hope this will not be done without
exhaustive research proving that they
have been wrong all these years.
So, having dismissed
the pitch issue as a red herring, we
come to the actual quality. I have compared
several key points, also using an LP
pressing of highlights from the performance
issued by EMI Italiana. In quiet passages
the LP proved to have the most lifelike
sound, most closely approached by the
EMI transfer, whereas the Naxos produces
a leaner, drier sound, just a shade
muffled. However, the LP was quite unable
to cope with heavy moments like the
Te Deum, which distorted badly. Obert-Thorn
obviously had to work from the LP but,
armed with better copies than mine,
and no doubt far superior equipment
to play it on, he succeeded in reducing
the distortion to an inoffensive level.
However, the EMI transfer shows us that
the distortion was not on the master
tape – the passage reproduces well.
So all-in-all, it looks as though EMI
still hold the trumps. Not to the extent
that you should worry if you have bought
the Naxos, but the EMI is now marginally
preferable.
To return to another
issue raised by my original review:
EMI are unrepentant in their insistence
on di Stefano and de Sabata instead
of Di Stefano and De Sabata. I believe
Italian is the only language in which
the various "di"s and "von"s
etc are written with a capital letter,
but on the other hand you look up Ludwig
van Beethoven in the encyclopedia under
"B" not "V" and
people will think it rather odd if you
talk about "van Beethoven"
instead of just plain "Beethoven";
whereas you look up Victor De Sabata
under "D" not "S"
and you must not refer to him as plain
"Sabata", so the capital letter
is there for a reason.
I’ve concentrated entirely
on technical matters, so let me close
by reminding you that this is one of
the very greatest recorded performances
of an opera ever made, so make sure
you get it one way or the other. Both
companies agree that you don’t deserve
a libretto at this price, just a synopsis;
but both synopses are good and EMI’s
note is a particularly informative and
well-written affair by Richard Osborne.
Christopher Howell
see review of
the Naxos
reissue by Robert Farr
Christopher
Howell's original discussion of this
recording with comments from Mark Obert-Thorn
Robert E. Seletsky
correspondence
received
Dear Mr. Howell,
A few more comments, I'm afraid. EMI
5 85644 2 is a budget version of the
GROTC mastering; it thus also begins
quite flat. Apparently, the speed drifts
during play, its pitch overlapping for
a bit with Mark Obert-Thorn's Naxos
version from the old LPs. The difference
is that Mr. Obert-Thorn tried to start
at a=440 Hz and painstakingly hold it
as steady as possible throughout. Perhaps,
in your estimation, Mr. Obert-Thorn's
efforts were wasted, but they certainly
demonstrate real professionalism and
enough care for an important musical
document to produce a version without
such significant speed drift as in GROTC.
The EMI Italiana LPs to which you referred
are in fake stereo, so they should not
be used as comparisons for anything.
All the fake stereo Callas LPs of the
1970s and early 1980s have enormous
distortion, probably because part of
the fakery was re-recording the channels
out of phase with each other.
Speaking of distortion and accuracy,
I should say that EMI is perfectly capable
of producing an accurate and compelling
CD TOSCA, and one without any pitch
problems. Listen to EMI 5 56304--the
full-priced Callas Edition version (pressed
in Holland), remastered in 1997 and
corrected in 1999 (after we called EMI's
attention to a nasty little problem).
It sounds like the best LPs without
compression, and there's no extra resonance
or strange stuff. As this version and
the best mono LPs have much in common,
it certainly must be a better representation
of the original tapes than GROTC. Moreover,
it was mastered by the same man, Allan
Ramsay, who apparently decided to create
a modernized sound for his GROTC version
in 2002. Further, the very first EMI
CD version, 7 47175, remastered in 1984
(perhaps by the late Keith Hardwick),
was also just fine, if perhaps a trifle
reserved. The fact that the unnecessarily
enhanced GROTC version and its budget
clone are now the most widely available
EMI versions is really the justification
for the existence of the Naxos edition.
--Robert E. Seletsky ("Callas
at EMI: Remastering & Perception,"
The Opera Quarterly, Summer 2000 &
"The Performance Practise of Maria
Callas," TOQ, Autumn 2004 [in press])
.................................
Alas, you have not dealt with my main
contention that "Vissi d'arte"
at least plays at 444hz not 440 in BOTH
versions, so that Mr. Obert-Thorn seems
to have transferred the recording SHARP
(and the EMI one drifts sharp as it
goes on).
Maybe the pitch of the performance
WAS 444 (although Italy should have
respected 440 at that date) so perhaps
the first thing to do is to establish
the likely pitch at La Scala in 1953.
Unfortunately my own enquiries have
only produced vague assurances that
it "ought" to have been 440.
Christopher Howell