Maria Callas LIVE at Last

The most eagerly anticipated opera release of 2017, Warner's set of 20 remastered complete live opera performances and five filmed recitals by soprano Maria Callas (1923–1977), hit the real and virtual stands on September 15. Commemorating the 40th anniversary of the soprano's death, which occurred on September 16, the physical edition of Maria Callas LIVE: Remastered Live Recordings 1949–1964 includes a 200-page, multi-language booklet that frames introductions to each opera with numerous photographs of Callas both rare and frequently reproduced. Each of the 20 opera sets within the box also boasts a handsomely reproduced front photo, usually from the performance.

As I write this, the box set is already back-ordered at but available at a higher price through In addition, online services are scrambling to post links to the 800 plus 24/44.1 remastered audio tracks in the 20 audio-only recordings. You can find HDTracks' collection here.

Maria Callas LIVE: Remastered Live Recordings 1949–1964's first three operas are the 1949 Nabucco, 1950 Parsifal, and 1951 I Vespri Siciliani. Each offers the only available "complete" recording of Callas in those operas. The set culminates with the 1955 Berlin Lucia di Lammermoor, conducted by Herbert von Karajan, 1960 Poliuto, and 1964 Tosca, followed by Blu-rays of the 5 concerts: Paris 1958, Hamburg 1959 & 1962, and London 1962 & 1964. All told, Callas never made commercial recordings of 12 of the 20 operas in the set.

Although virtually all the set's audio-only performances have been available in whole or part in various pirate editions, Warner claims to have uncovered better sources in some instances. In addition, Mastering Engineer Christophe Hénault at Studio Art et Son is credited with significant pitch correction, hiss reduction, etc.

Before discussing the sound, one question must be addressed. Why is Warner devoting so much energy to remastering often dimly recorded live performances by a soprano who died 50 years ago, and whose prime amounted to a mere 10 years (1949–1959)? Simply put, Maria Callas changed our understanding of opera. Not only was she responsible for the bel canto revival that brought us Rossini's Armida and Donizetti's Anna Bolena, but she did so via a remarkable, anything-but-conventional voice that could encompass both florid, coloratura roles and heavier, more dramatic assumptions. When both Callas's light, flexible voice and heavier emissions came together, and she moved between pathetic utterings and cries of fury, as they did in Bellini's Norma, Donizetti's Anna Bolena, and Verdi's Nabucco, Macbeth, I Vespri Siciliani, Rigoletto, and La Traviata—all of which are included in this set—the results were spectacular.

Callas's voice was anything but conventional. On record, only the 19th century diva Lilli Lehmann also proved herself fully capable of convincingly singing both Wagnerian drama and Mozartian high coloratura. Callas's low tones, especially her stone-cold chest voice, were those of a mezzo, while the highs seemed from another voice entirely. Those highs, it must be noted, were often steely. In the opera house, they had cutting power; on recordings, you may sometimes wish that the microphone had been placed farther away.

In addition, parts of Callas's midrange were often covered, with diction a bit murky. Throughout the range, the sound rarely approached anything we could call sweet or "pretty." But the voice was nonetheless remarkable for its power, flexibility, and ability to convey emotion.

Of equal import is Callas's extraordinary emotional veracity and virtually flawless sense of timing. Tragic emotions seemed allied to her being, with pain and suffering central to the voice. Anger and rage, too, burned at her core. Beyond the roles in which she was famed for her fury, her early Aidas—the 1951 Mexico City performance in which she interpolated an astounding E-flat at the end of the Triumphal Scene is in the box—include renditions of "Ritorna vincitor!" so forceful that you can be sure anyone within earshot was quaking at the thought of anything less than total victory.

She also took risks. In her earlier performances, she routinely interpolated forceful high Cs, Ds, and E flats. She also tended to push her dramatically dark low tones up past the natural range break and into the middle voice. All this, along with a fiery personality that helped provoke one scandal after another, took their toll on her voice. Those who try to pretend that Callas's emotional life did not affect her voice, and that the voice is not a window to the personality and the soul, are proponents of bifurcated thinking that ignores the mind-body-spirit connection.

As early as 1955 (Lucia), when Callas was 31, the occasional wobble intruded on otherwise near-flawless vocal production. A few years later, high notes began to erode, and the wobble moved a bit lower into the range. By the end, the highest notes were gone, notes above the stave were frayed, and breaks between registers were painfully exposed. When Callas sang her final operatic role onstage, she had not yet turned 42.

In the midst of her infamous 1958 Dallas performance of Cherubini's Medea, Callas received a telegram from Rudolf Bing, General Director of the Metropolitan Opera, firing her from the Met. Railing to reporters who were filming her in her dressing room, she claimed that she had refused to sign her Met contract for the next season because Bing kept offering her the same old roles, which he wanted her to sing over and over again during the season. Truth be told, 1958 was the last year that she dared attempt onstage the high notes in some of those roles.

The new set includes the 1953 La Scala performance of Medea, with Bernstein conducting, that shows Callas in complete command. Tosca was one of the few roles that Callas was able to perform convincingly until the end. Despite 5 high Cs, and a famed aria ("Vissi d'arte") that climaxes on a sustained B-flat, Tosca allowed her to triumph even as other roles fell by the wayside. The final live opera recording in the set, the 1964 ROH London Tosca conducted by Cillario, is an invaluable supplement to the indispensable staged Act II Tosca that is included on the set's third Blu-ray. For those who wish to understand what the Callas mystique was about, viewing this performance, which also includes Tito Gobbi as the most delicious, aristocratic sadist imaginable, is essential.

The Sonics
Many of Hénault's remasterings come complete with disclaimers. Opening the box of the earliest performance, Verdi's Nabucco from Teatro San Carlo di Napoli conducted by Vittorio Gui, one reads, "This 1949 recording has come to us in a very poor state on account of its age and the recording techniques used. We have nevertheless decided to publish this document . . . because the mint source we were able to use offers despite its faults a quality and richness of sound particularly in the highs, far surpassing any of the CD releases with which we compared it. We are nevertheless aware that numerous instances of saturation and noise may make listening difficult, particularly in Act IV."

You can say that again. Opera aficionados hoping for miracles—I'm one of those—would be wise to adjust expectations. You can do only so much with a severely compromised source, of which there are many in this set.

Jumping ahead to the 1954 La Scala set of Gluck's Alceste, we read, "Unfortunately, this Alceste comes to us with very problematic sound quality. Retuned tape pitch and a light equalization cannot mask the recording's flaws, aggravated further by the ever-present prompter. We nevertheless took the decision to release it, on account of Giulini's splendid conducting and, of course, the historical interest of Maria Callas's sole Alceste."

That said, comparing earlier pirate issues of some of the operas with Warner's remastered offerings find the improvements readily apparent. Take, for example, the rare, startling 1952 Florence May Festival performance of Rossini's Armida, in which Callas's rendition of the 7-minute aria, "D'Amore al dolce impero," includes three interpolated high Ds. The last of these is so astoundingly large that it inspires delirium among audience members and listeners alike. (Count me among them.) In comparison to the Melodram issue, Warner's new remastering offers what feels like a 100% improvement in color saturation and clarity. Although "low-fi" may be too kind a descriptor when discussing this and some of the other pirates, what Warner presents is far more listenable than before.

The famed 1955 Lucia, featuring Callas in limpid voice, offers comparable improvements. The singing is extraordinary. Callas's rapid shifts between two of her many voices—Callas at her most pathetic and vulnerable and, at the other extreme, Callas at her most agitated - delineates Lucia's madness better than virtually any soprano on record. The brilliant voice of Joan Sutherland, whose debut in Lucia I saw at the Met, may be able execute coloratura faster, but her command of erratic emotional shifts is minimal. As fine as the beloved Sills may be in the role, it is ultimately Callas whose pain will tear you apart.

Shortly before the UK's Pearl label went out of business, it issued a remarkably quiet recording of Callas's fabled 1958 Lisbon La Traviata in vivid sound. Warner's remastering lends more body to the voice, at the loss of some transparency. It's a toss-up as to which is ultimately superior. But this is probably the only toss-up in the entire set.

As to the question, "Why not 24/96 rather than 24/44.1?" I can only guess: Either the studio was not equipped for higher resolution (doubtful), or the source material was not deemed worthy of same. It's important to note, however, that many of Sony's recent hi-rez downloads of newly recorded material are similarly limited to 24/48. Might it be possible that, if the remasterings we actually made at higher resolution, Warner is saving them for future MQA release?

Regardless, while I was unable to audition any of the hi-rez downloads prior to the posting of this review, experience indicates that the leap from 16 to 24 bits delivers truer, more saturated colors and larger, more lifelike images. Every increase in resolution gets you closer to the source. In the case of what Callas offered in her live performances, this amounts to a lot.

For sweet love songs, lyrical excursions, and art song, you'll want someone else. For love wounded, love enraged, vengeance with a capital "V," and coloratura with substance, Callas is essential. Even when the voice does not seem a perfect fit for the character at hand, Callas's musical and emotional intelligence will convince you that her understanding of what composers and librettists wished to express is unassailable.

In either physical or download set, every opera lover must have this set. Some, understandably, will balk at the sonics, and instead opt for commercial recordings of the set's eight operas that Callas recorded in studio. But it was onstage that Callas held nothing back. No other singer on record can match her intensity, commitment, naked emotional truth, and audacity. Maria Callas LIVE: Remastered Live Recordings 1949–1964 is indispensable.

dalethorn's picture

Funny thing, that to sample my way through this Callas set, I need to clear some spaces on my computers and music players. Having purchased an 'audiobook' by a former first lady containing 14 CDs, I'm asking myself "For spoken word reproduction, do I need the full 44 khz bandwith in my CD rips? What if there are some sounds that actually need a higher resolution?" And now that HDTracks and others sell 96-192 khz downloads of mono recordings from (in some cases) 78 RPM discs, not a simple question.

tonykaz's picture

Well, of course we should.

Geez, it's less than $5 per disc. - 22 discs. for the entire grouping.

Most everything I own of Callas is from "the Best of" type Albums, about 25 5minute pieces, dating from 1953 - 1965, recording quality not so good, at times.

This is a good example of why Digital is important: How could we ever have this set in Vinyl?

Thanks JVS

Tony in Michigan

ps. I'm itching to hear more about those Amps

Jason Victor Serinus's picture

I will leave to others the big question, "Does it make sense to buy vinyl transfers of digital masterings and remasterings?" Not having a vinyl rig due to space and other considerations, I'd just be repeating what true experts such as Michael Fremer have to say on the matter.

Regardless, the market would have to demonstrate sufficient interest to make Warner's investment in vinyl transfers worthwhile. IMHO, if you can forego the booklet essay and photos - I expect some of the information is in PDFs that accompany the downloads, although I'll know more about that in a few days -and are equipped for file playback, the 24/44.1 files are the way to go. They will certainly sound better than the original vinyl issues of a few of these performances, which were done without adequate pitch correction etc...

tonykaz's picture

Vinyl would cost in the range of $800 figuring $38 each type prices.

Could we get Kassem interested?, I doubt it.

Would I finance it ?, no way!

Would it sell against this CD Set?, hmm, no!, not even to Vinyl guys.

It probably needed a ton of 'digital-cleaning-up' to make presentable
and even then it's mostly of Historical Interest for old geezers like me who had real time discussions about Callis with people that knew her. ( of course my Mom admired her ).

We're lucky that somebody took the time and effort to do this set.

Tony in Michigan

ps. a proper Vinyl rig would cost in the $$,$$$ range, well on the high side. The Storage System would need a 2,000 Cubic Foot commitment, not to mention the huge cost of the vinyls themselves. Vinyl is a Collector's Hobby in-of-itself, it dominates one's Life.

ps.2 I think that folks like Callis could sing to the back seats, I'll bet that they have dynamic range in excess of 16 Bits and well into 20 Bits. What Recording System ( did anyone ever have ) that could capture an Opera Voice's full performance? Opera is strictly in the 24 Bit range. Isn't it?

Jason Victor Serinus's picture

I highly doubt Warner is interested in giving its brand new remasterings to a specialty audiophile label.

A well-versed recording engineer could address the "sufficient bits" question far better than I. It is my understanding that anything with wide dynamics, undertones and overtones, and multiple octaves demands at least 20-bits, with 24 far more satisfactory.

Certainly, pirate recordings made from broadcast are highly compressed. So are most pirate recordings made with recorders hidden in the audience.

It is also doubtful that Callas's studio recordings were made without compression. As John Atkinson pointed out to me, if you watch the video of the making of the Solti Ring, you'll see the engineers riding the master volume control as Flagstad, Nilsson, Hotter, and others are singing. They were concerned with tape saturation, as well as dynamics so great that needles would jump out of the groove and people would be constantly jumping up and down to adjust volume.

tonykaz's picture

It's been disappointing to hear recordings after hearing "Live".

With 24bit depth we stand a chance of having a Full Voice but Bob Katz says that Recording engineers will only get close to 22 Bit depth because going over 24 sounds horrible. 22 Bits still yealds 22 x 6 = 132 Db of dynamic range, doesn't it?

Chad Kassem's people are say'n that they just bought a 46,000 Classical Album Collection. Hmm, can we imagine the Storage Requirements for a Collection that large? Kassem himself claims to own a 20,000 personal Collection, phew! Todd the Vinyl Junkie says he owns 12,000.

Acoustic Sounds actually say that they are servicing the "Collector" and encouraging the young, new Collector. It's a collector type business.

I guess that I'm not a Collector, I want the 'mood-altering' that music creates, for me it's a Drug of Choice.

Viva la Music,

Tony in Michigan

Robin Landseadel's picture

I doubt I'm the most well-versed audio engineer, but having heard/owned recordings like those found in this set—lo-fi 50's classical music—I can say that hi-def is overkill for this sort of source material. These are often air checks, limited in both frequency range and dynamics and loaded with both self-noise and noises off. Note on those "Ring" recordings—recording engineers riding gain is an indication that they are working through the self-noise of their gear. With 24 bit recording one sets the level for the peak and then compresses in post-production, much easier to to get consistent results. In any case, a good MP3 ought to capture everything that's in these recordings. MInd you, the real trick isn't in bit depth or frequency width but in simply sounding good. I'll take a good sounding [i.e. well mastered] MP3 over lo-fi 24/192 FLAC any day.

foxhall's picture

I'm very cynical but I hope the decision to release 24/44 was, in fact, because a higher sampling rate didn't offer any difference.

I've heard "experts" claim that bit depth is far more important than sampling and other "experts" claim sampling rate is more important. The argument gets old fast.

Jason Victor Serinus's picture

People love to create false divisions around this vs. that so they can argue. I agree; it, and countless other audiophile "debates" about which is superior or what's most important do get old. Better to simply listen for yourself.

In my own listening tests, I've found both bit depth and sampling rate important. I have confirmed, however, that simply increasing bit-depth without increasing sampling rate does make for a far more convincing presentation.

Anton's picture

This will be a great 'academic' chance to listen for 'the decline' that people discuss.

I always found her to exemplify 'brooding' in her characters, which I mean in a good way.

Jason Victor Serinus's picture

The Callas Live albums are already available for streaming on Tidal in MQA.