The New Maria Callas Remasters: Good or Bad?

The other day, I posted about the new box set of Maria Callas’ Complete Studio Recordings being available for download on the iTunes Store. I had a few exchanges with Andrew Rose, of Pristine Classical, which restores historical recordings, and Andrew said that he thought the Callas remasters were not good. He told me he was writing something for his newsletter, and granted me permission to reproduce it here.

Here’s what Andrew Rose has to say about these remasters.


There’s been a surprising amount of fuss about a new Maria Callas box set recently. Music and tech blogger and Macworld writer Kirk McElhearn noted its appearance on iTunes – “This is the first big classical box set I’ve seen on iTunes sold as a set” he wrote on Facebook this week. I’ve also seen it popping up on music websites – on Qobuz, for example, the music is being promoted with a picture of a box set, but it’s actually being offered across individual albums in various formats up to and including ultra-hi-resolution 24-bit 96kHz lossless downloads.

A couple days ago I also received an e-mail on the subject from a Pristine Classical newsletter subscriber called Rob (who I’m sure is reading this), who wrote:

“Since you’ve tackled a few, when you get a spare moment, hoping to hear your appraisal of Maria Callas Remastered (The Complete Studio Recordings 1949-1969).”

I clicked on the link in Rob’s e-mail, which took me to Amazon’s UK website where the new set could be purchased on CD for £210.99. Bizarrely the same set can be bought on Amazon’s French website for €199 – which is nice because Amazon currently reckons £210.99 equals €280.34!!! Eighty Euros is quite a saving – if you’re in the UK and want to get the set check out Amazon’s European websites first.

Anyway, it so happens that I already own, alongside a stack of original vinyl Callas albums, EMIs’ last Complete Studio Recordings box set, released in 2007, which looks a lot like this:

85.png

It’s also considerably cheaper right now – £77.38 (€108.81) on Amazon UK, or better still, €69.90 at Amazon France. (Tip: French buyers get cheaper Callas, it seems – the UK EMI box is 55% more expensive.) That’s a third of the price of the new edition for French customers. So is the new remastered version three times better than EMI’s old version?

(As an aside, a friend of mine with a vineyard in St. Emilion once remarked that Chateau Petrus, one of the world’s most expensive wines, was made a mile or two down the road from his own, but cost 50 times more per bottle than his did. Was it better? Yes. Would he rather have one bottle of Petrus or fifty bottles of his own excellent Grand Cru? The latter, please…)

Sorry for the interruption – back to my question: is it worth, as a listener, paying three times as much for the new remastered set? In a word, “no”, and I won’t be investing my money in it. What I have done though is invested in some of the ultra-high-resolution download tracks – so you don’t need to – and subjected some random samples to a few checks and tests. This also allowed me to follow up the (inevitably) speculative response I gave to an e-mail I received back in August from the music critic David Patrick Stearns:

“As you’ve no doubt heard, Warner/EMI is bringing out a Callas box with new remasterings taken from master tapes and I’m doing a story about it for the Philadelphia Inquirer.

To be honest with you, I like your remasterings better. It seems to me that the difference is this: Warner/EMI is trying to bring out all that’s there on the master tape, the odd result being that Callas voice has rarely seemed so harsh to me.”

So, what exactly do you get for your money? My expectations were admittedly low, and in my first test I wasn’t surprised with the results. Yes, it’s been transferred this time at 24-bits and 96kHz. No, it doesn’t sound any different to the older EMI issue that’s now a third of the price. (This was a track taken at random from a 1955 recording of Madame Butterfly – you’ll be able hear my own take on this recording in a few weeks).

There was only one audible difference between the two: the new one was louder. Yes, turn your amp up around 5dB when playing the old transfer and you can have all the joy of the new one!

I concocted a track which began with the older transfer, cross-faded into the new version, then cross-faded back to the older one. With the levels matched I listened over and over again, on my studio monitors and on my headphones. I couldn’t hear the joins and I couldn’t detect any qualitative difference whatsoever. There was nothing at all to indicate that I was listening to a “new and improved” recording, however hard I listened, and I like to think my hearing is reasonably perceptive, given my line of work.

The 24-bit transfers have clearly helped the engineers to maximise the sound levels (they can set record levels a bit lower and then increase the final levels digitally to get the most out of a CD’s 16-bit sound without bringing up any audible digital noise).

But what about those super-duper 96kHz sampled high frequencies? There’s quite literally nothing there.

To be technically correct there is of course something there: a whole 26kHz of uninterrupted tape hiss from 22k up to 48k. Tape hiss, being random low level noise, has no practical frequency limit – if your tape heads can reproduce it it’ll probably go on up and up, almost ad infinitum. But those microphones, those amplifiers, those mixers and those tape record heads in use back in 1955 did not go on up ad infinitum. They started to roll off somewhere around 20kHz, and by 22kHz you’ve pretty much reached the upper limits of what was captured. Which is as high in frequency terms as a CD can go, as it happens. Above that there was no musical content whatsoever on the 1955 recording. So there’s no advantage to be had whatsoever in wasting money, bandwidth and storage space on a 96kHz version of the recording in question. There’s plenty of music contained within the constraints of a standard 44.1kHz CD. But beyond that, the rest is noise.

Next I turned the clock back to 1953 and one of the most famous Maria Callas recordings of them all, Tosca – one of the first Callas recordings I tackled here at Pristine (PACO 080). This time there were minor differences to be heard over the previous outing on EMI. The pitch on the new transfer was marginally higher (closer to my own pitching than heard on the older transfer, though they’re all close). There was also what appeared to be a slight boost in the treble.

But what was most interesting here was the new dynamic range – or rather the lack of it. With the loudest moments set up to match between the two transfers (leaving my own out of the picture for the time being) it was noticeable that the new transfer was louder in all the quieter sections. This suggests one immediate cause: compression of the audio signal to make everything appear louder. Not in the way Rock music does it – this is certainly subtle – but the irony here is that this new 24-bit transfer, which as such has much greater potential dynamic range, uses less of that range that the old 16-bit transfer did. By compressing the audio levels (and no, this isn’t data compression as in MP3, don’t be misled by the terminology), essentially squashing the dynamic range, this might (along with the equalisation) be part of an answer as to why my Mr Stearns felt that the Callas voice had “rarely seemed so harsh to me”.

On the plus side for the new set? There’s no denying the box set is lavishly packaged. So much so, one can’t help but wonder whether this is where the budget for this project went, rather than spending lots of time and effort in the remastering suite (it doesn’t take a genius to load up a tape machine and press play and record, after all). I may be wrong of course – this is speculation based on sampling two discs out of 69 – but it’s informed speculation, it’s measured speculation, and its tested speculation.

In making at least one recording subjectively “louder”, by reducing its dynamic range, one might consider that the new release is in this respect actually inferior to its predecessor, still available at a third of the price. It certainly isn’t using its full 24-bits, nor its 96kHz sampling rate (though it’s possible that the later recordings might offer the odd cymbal crash the breaks the upper limits of the CD). And if other recordings simply invite you to turn up the volume a small amount in order to replicate the “new” sound, why buy the whole lot again – or rather, why not go for the cheaper set for as long as it lasts?

If packaging is your thing, or you want to splash your cash and get the “Mastered for iTunes” lossy download – at €149.99 here in France, still more than double the price of the 2007 EMI non-lossy CDs – then why not blow a couple of hundred on the set? It’ll look very nice on your shelf. But don’t throw away your old copies just yet, whoever remastered them…

Andrew Rose


It’s interesting to note that Sterophile raves about these remasters. The American magazine says:

“Once we get to material remastered from the analog tape masters, the restoration of the treble edge of Callas’ voice and high-frequency extension of the orchestra, combined with palpably greater richness and body, sounds as though a previously clouded window on Callas’ artistry has been scrubbed clean.”

So who to believe? I know that Andrew has not only listened to the recordings, but also analyzed the sound, measuring such things as dynamic range, and the Stereophile writer says nothing about that. They did listen to some of the new remasters compared to the old, but, what a surprise, they didn’t choose what they heard.

“Warner supplied members of the press with a CD that compares the last 16/44.1 remastering of five arias with the new 24/96 versions downsampled to 16/44.1.”

So, Warner chose the ones that would sound best compared to older versions… Pretty sly marketing strategy.

If you’re interested, you can get this set from Amazon.com, Amazon UK, Amazon FR, or the iTunes Store.

And if you like historical recordings, stop by Pristine Classical and see what they have to offer.

5 thoughts on “The New Maria Callas Remasters: Good or Bad?

  1. Having written at very great length in The Opera Quarterly about Callas’ recordings in articles published between 2000 and 2005, I comment on the Warner remasters, if very belatedly. Mr. Rose has his own agenda, with which I personally take issue. He is marketing from-LP, “ambient stereo” versions of the Callas studio material. That is his right, though I find them neither appealing nor accurate. Asking him to rate his competition cannot yield unbiased results.The conclusions he draws are disingenuous as, by his own admission, he has only listened and compared a few discs. I agree that the Warner set is mastered louder–too loud in a number of cases, actually accentuating the overload distortion present in the old tapes, or creating new distortion–however difficult the latter is to believe, given the digital format: the ‘harshness’ that shows up in Callas’ voice from time to time, which we have not heard before, as well as occasional over-present bass. To say, however, that the 1997-2005 EMI versions and the new masters sound identical is bizarre. At the same volume levels, one can hear very clearly that the Callas voice in the new version regains much of the creaminess and immediacy only heard on the best of the original LPs, qualities absent from the 1997-2005 EMI offerings. Moreover, one must note that the Cetra material is also included and has never sounded as beautiful or as faithful to the original LPs. A significant point is that many, and quite ghastly, new editing and pitch errors were introduced into the 1997-2005 EMI CDs, and are gone in the Warner set. As important to note is that those 1997-2005 releases were simply re-EQs of the Keith Hardwick 1985-89 transfers, the original CDs virtually all superior to the 1997-2005 re-EQs and lacking in editing and pitch errors. If anything, it is those early CDs which are closest to the spirit of the LPs in many cases. However, being unavailable for a long time, the Warner set, which at this time, is far less expensive than anything that preceded it–as little as $3 USD per disc–is the only option, the 1997-2005 error-ridden, poor-sounding versions happily gone.

  2. I have to agree with Robert that I believe Andrew Rose is being ever so slightly disingenuous. In recent years he has remastered 10 Callas operas and offered these for sale through his company Pristine Classical. So there is certainly a conflict of interest involved.

    My fascination with Maria Callas started in the mid-1980s with the acquisition of a large vinyl collection of the beloved diva. As ever with vinyl, some sounded fine, some most certainly didn’t. So I started buying the EMI CDs. Bearing in mind the period, these were the first CD remasters from the late 1980s – as described by Robert. Transfers by Keith Hardwick I now understand. Ever a fanatic for sound quality, I have to confess I found the great majority of these CDs to be more than acceptable to my ears. These were the attractive EMI boxsets with glossy black card boxes – I still treasure these. But then EMI lost it’s way, and “ART” was foisted upon us. The next generation of Callas remasters were so inferior as to send many collectors in search of the earlier glossy black boxsets. I could write a book about the failings of these 1997-2005 releases. I will just say that they sounded digital in the worst way – the sound lacking in bloom, tonal colour, ambiance, etc etc. Additionally there were some awful cock-ups – for example, Walter Legge’s painstaking creation of Tosca’s entrance lost in the digital destruction.

    So I am also puzzled why Andrew hears so little difference between his 2007 boxset and the new remasters. I haven’t yet purchased the “Maria Callas Remastered” boxset in question, however I have bought two of the Warner remastered individual sets – “I Puritani” and the 1953 “Tosca”. I also own the first generation glossy black EMI boxsets of these two operas, I also have downloads of Pristine Classical’s recent remasters. I wouldn’t dream of an all-out attack on Andrew Rose’s remasters, because in their way they sound very enjoyable – they at least sound more natural than some of EMI’s efforts. However, I feel Andrew’s remasters lack real transparency – it’s as if the music has been a little “smoothed over”, certainly leading edges are slightly rounded, blurred if you prefer. Perhaps most significantly, “ambient stereo” is no replacement for those subtle acoustic cues which exist even on 1950s mono recordings.

    I will go ahead and purchase the 69 CD boxset and I’ll hear for myself whether Callas sounds harsher, whether low level detail has been artificially boosted. Or in fact whether a further veil has been lifted from the music. I will say that the two Warner sets I’ve purchased individually are the best remasters I’ve ever heard.

    • I must apologize for allowing myself to be misled. As of November 2016, following two years of exhaustive listening tests and spectral analyses, I’m afraid I must agree with Andrew Rose (whose own work is still not to my taste): Warner’s ‘Callas Remastered’ does not seem to be a new mastering at all, but a retread of EMI’s 1997 ‘Callas Edition’ and 2002-5 ‘Great Recordings of the Century’–themselves re-EQs, by the same engineers who produced ‘Callas Remastered’, of Keith Hardwick’s first, and still best, 1984-9 digitizations.

      The timings of the 1997-2005 EMI discs, with rare exception, are identical, *to the second* with those in Warner’s version by the same team when one subtracts Warner’s longer silences at the beginnings and endings of discs and acts–a physical impossibility for something supposedly redigitized from analog scratch. Moreover, the sound has nothing to do with the analog sound. It is re-EQed to be a harsh and boomy version of the ‘Callas Edition’, very fatiguing to hear, but with the same recessed voices. The LPs and the aesthetic of mono operatic recording in the 1950s–especially at EMI–show the forward placement of the voices with instrumental accompaniment, not the other way around, crisply but never harshly, presented. Spectral analyses shows the unfortunately chosen re-EQing.

      Rose is right that, certainly in old mono recordings, there is nothing above 22kHz, making the 24bit/96kHz ‘HD’ remastering a red herring. That Warner claims 69 discs worth of material have been remastered from the analog tapes in such a short time is impossible to believe when it took Keith Hardwick five years. ‘Callas Remastered’ is apparently the ‘Callas *Edition* Remastered’. One should also bear in mind that in 1997, the only recordings claimed to have been remastered from analog sources were the recitals, but the whole series was done at 20/88.2kHz. So any stereo material with frequencies higher than 22kHz (and there is some) would already have been attended to; the 24/96kHz is cosmetic.

      As for ‘original’ analog tapes, that too is a red herring, as the tapes were copied continuously for decades to send to various international EMI branches for the creation of new LP stampers. If the Abbey Road people used analog sources–and I do not believe they did in 2014–it would hardly have been ‘original’: it would be whatever tapes were stored in the UK. And surely, the Cetra material would lack original tapes, having circulated in the public domain for half a century. Warner owns Fonit Cetra as well as EMI, and the Cetra operas are most likely re-EQs of Cetra CDC-2 TRAVIATA and CDC-9 GIOCONDA, the 1949 arias CEDARized versions of Warner’s own 1999 ‘Il primo disco’. Photos in the book accompanying the Warner set show purported ‘original’ tapes; but we must remember that most recordings were made in Italy and France, yet the tape boxes seen are from American manufacturers, and might be anything.

      Head engineer Ramsay (of Callas Edition notoriety) takes misplaced pride in revealing the bizarre addition of faked record roar between bands in the recitals and at the beginnings of Acts and discs, to replicate LP sound; he seems to be unaware that those of us who bought the LPs would have returned discs with this characteristic as defective. No other artist’s work has been subjected to this odd treatment on CD. Ramsay makes it sound like a new idea, whereas it was present in the 1997 edition as well. At least the errors, introduced in 1997 to Hardwick’s initial digitizations by the same engineers–to which I called public attention in The Opera Quarterly and elsewhere–have finally been corrected, but that’s really all the best this set has to offer. Warner evidently spent its money on hype, not real remastering; ironically, these discs are sold for super-budget prices, without the great essays by John Steane and minimal notes by others, and libretti on CD-ROM only if one buys the whole set. If I am wrong, it really shows these engineers in a very poor light. I still recommend the 1984-9 EMI/Angel digitizations by Keith Hardwick (which often sound better in EMI Classics pressings of 1991-4) if one cannot find, or has no means to play, the LPs.

  3. Andrew Rose is trying to sell his own Callas remasterings, and if he tells us the circumstances surrounding the advantage of Warner having access to the “original” tapes is compromised by their not being actually original, this is to deflect attention from the fact that HE has no access to master materials. Quite often he is not forthcoming on what his sources are at all, so they could just be reworked from commercial CD releases OR LP pressings…who knows. If some hear his fake stereo and overprocessed versions as desirable, so be it, but they are no substitute for any of EMI/Warner’s versions, which to varying degrees represent original materials, not pseudo-audiophilia tweaks and falsifications. I’m not saying I’m crazy about these new Warner masterings. They improve on the previous collected set, but the first EMI cds are in some cases still the best digital versions. A more honest approach to the kind of work that Rose does is represented by Mark Obert-Thorn’s transfers for Naxos Historical, which are taken from good LP pressings, if you want the closest to the originally released versions. In the case of “I Pagliacci”, one really SHOULD avoid the previous EMI cds and opt for the newest Warner version, as it is the only issue to date without added reverb.

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.