Chapter 9
Recordings vs. Performances (vs. Recordings of Performances)
Imagine, for a moment, a world without recordings. In order for you to hear any music, you would have to play it yourself or attend a concert. This would be a world in which a Beethoven symphony was an event. Experiencing violin virtuoso Niccolò Paganini required a journey to be made, a commitment of time and attention. It was both social and societal. If your family was rich enough to own a box at the opera, it was about seeing and hearing as well as being seen and overhearing.
All music was contemporary. With the advancement of railroads in Europe, one could journey to a little town in Bavaria to hear Wagner’s latest opera, Parsifal, for example, as the young Debussy, Stravinsky, and Ravel did. If you were Henry Ford, you took an ocean liner and the train to Munich in September of 1910 to attend the world premiere of a new symphony by Gustav Mahler. Other visitors that night included King Albert I of Belgium, Georges Clémenceau—the once and future prime minister of France—composers Richard Strauss, Anton Webern, and Erich Wolfgang Korngold, and the writers Thomas Mann and Stefan Zweig. Music and the performance of it were synonymous and constituted an event.
All that had begun to change in 1877 with Thomas Edison and his cylinder machine. Once it was commercialized a decade later and the all-wax cylinder became the standard format for “records,” music came into people’s homes that had been “recorded.” Ricordare is Latin for “remember,” “commemorate,” “recollect.”
At approximately the same time, Marconi’s “wireless” of the 1890s was being developed and expanded into broadcasting music on telephone lines. In 1911 Marcel Proust subscribed to France’s Théâtrophone, which was linked to eight Paris theaters and broadcast live performances that he heard via headphones in his bedroom, where he was able to hear Wagner’s Die Meistersinger from the Opéra and Debussy’s Pelléas et Mélisande from the Opéra-Comique. However, broadcast technology did not allow for repeated hearings, the way Edison’s records could.
While you might know that Edison’s cylinder machine was replaced by discs by the 1920s, you might not know that the personal cylinder machine was also capable of recording. Making use of the same playback mechanism, it was possible to record your own voice, your own musicmaking, or someone else’s performance—and repeat it. When the Metropolitan Opera’s librarian Lionel Mapleson took his Edison cylinder machine and placed it in the prompter’s box in 1901, he made it possible to hear the greatest singers of the age in performance.
Equally significant was the fact that each time you played a cylinder you simultaneously degraded it. The grooves in the wax were gradually worn away by the stylus that vibrated and sent its analog information to the cone out of which you heard music. Each hearing was less good than the previous one, ultimately leaving us today with the last echoes of that era; the more beloved the records, the more distant they became. They represent a slow-motion and inevitable fade to background noise, over which we hear a phrase, a high note, and the crackling that gradually engulfed and ingested the art we desperately wished to remember and commemorate.
Today, music is only occasionally an event. Most often it is accessible everywhere and at any time because of recordings. It is safe to say that more people are hearing music today than ever before in history. While you may have heard Beethoven’s Symphony no. 5 hundreds of times, how many of them were live? Can Beethoven’s Fifth ever be an event for you? (Answer: yes, but only under certain very particular circumstances.) Certain works—Mahler’s Eighth Symphony, Wagner’s Ring—remain events, but few works create the obvious need to attend, though one could argue that in order to truly experience music one has to be present for it.
Since music is invisible, ethereal, and evanescent, the development of sound recording has given the world something that has fundamentally changed the way we can experience it. It has also given conductors a surprising gift: something permanent to leave behind, documentary evidence of the notes, the text, and the point of view. It is a clouded window into the past, footprints of dinosaurs, insects in amber. It tells us so much but can never replace the process of a living performance, one that includes the unknown, the danger, and the participation of those present. Recordings are snapshots—processed, airbrushed, and Photoshopped pictures—and like all the metaphors I have just dredged up, recordings are in fact metaphors.
Perhaps we should look at recordings as a separate form of performance art, one that relates to living music but is quite different from it. When the microphones are capturing a live performance, the archival results differ significantly from the impact of attending that performance. If you had attended it, it inspires and preserves a cherished memory. If not, it acts as an imperfect aural manifestation of other people’s experiences.
The final moments of the recording of Wilhelm Furtwängler conducting Beethoven’s Ninth Symphony at the postwar reopening of the Bayreuth Festival is an orchestral mess, one that used to cause merriment among my friends when we were in high school. If, however, you were at that performance, no one in the theater heard any of that. No one was actually listening. Instead, all were participating in the rebirth of a great and holy shrine of German music, hearing an “Ode to Joy” that confronted the horrors, the rage, and the guilt in postwar Germany. The greatest living German conductor attacked the coda of a universally beloved masterwork that had emerged from the German soul a century before two world wars, and conducted it faster than anyone could possibly play it. The audience experienced a final, glorious, and defiant flash of lightning on that night of July 29, 1951. The archival recording tells a very different story, and we have no right to judge it with cool detachment, sitting in our living rooms three quarters of a century later and sniffing that it “isn’t together.”
How conductors use recordings is a fundamental development in the art, one that has changed it forever. On the one hand, it allows us to study the resultant performing choices of the second and third generations of maestros, those who emerged just after Wagner, Mendelssohn, and Berlioz had invented our field. On the other hand, it provides an opportunity to establish a legacy that defies the ephemeral nature of our art. Conductors would otherwise be reduced to fading memories and the reviews that no one remembers. One could of course argue that recordings have destroyed the purity of our art, which is, after all, the art whose outward manifestation is the waving of our hands in the air so that others can make music. There is no recording of a performance conducted by Mahler or Hans Richter, just as we cannot hear Liszt or Chopin play the piano. We begin to peek into the gaslight-and-candle era with the most distant and unearthly sounds—a snippet of a great opera singer, a whoosh of a piano that is being played by Brahms—like having a crystal radio set in the attic picking up emanations from the past. We can barely perceive the outlines of music and style. We want more, but there is no more.
Conductors may or may not make use of historical recordings in preparing to interpret music, whether learning Beethoven or Strauss or studying for a second performance of a new work. In preparing Leonard Bernstein’s A Quiet Place for its European premiere, our having archival recordings of John DeMain’s 1983 world-premiere performance with the Houston Grand Opera was essential in understanding the timings of the various scenes and dispassionately hearing what needed to be changed in its orchestration. The same was true of Menotti’s La Loca for its New York performances, after its 1979 world premiere in San Diego, which had been conducted by Calvin Simmons. In both cases, the composers made use of the recordings in order to make changes and give advice to this conductor, after the heat of the world premiere had cooled down.
That was a generation ago. Today, composers frequently create demos of new works if they are making use of contemporary engraving notational computer programs. These programs “read” the score and transform it into fairly realistic orchestral sounds. In the 1940s, as already noted, Koussevitzky employed two pianists to play new music for him. Today, a conductor can study the resultant sounds created through sampling systems that predict what the live orchestra will sound like before a world premiere takes place.
If we find ourselves going back in time to “study” with great interpreters, we must face certain facts. We have to know how the recording was made: For example, was it recorded directly onto master discs that were limited to a maximum of five minutes a segment? Did that determine the tempo, or did the maestro and his engineer simply chop the music into sections, after having timed a prerecorded version? In 1927, the great Wagnerian conductor Karl Muck refused to record the “Good Friday Music” from Parsifal in three chunks. As a result, the composer’s son, Siegfried, stepped up to the rostrum in the Bayreuth Festspielhaus and gave his interpretation of the music.
Muck subsequently thought better of his decision, and recorded this same music two years later in Berlin, perfectly matching his tempos so that today we can hear his recordings spliced together as a single entity. What makes this story so important to a contemporary conductor looking to perform an “historically informed” reading of Parsifal is how different these two conductors’ readings are. Muck was considered the greatest interpreter of Parsifal in the generation after its first conductor, Hermann Levi, who led its world premiere in 1882. The Wagner family so admired Muck’s readings that he conducted Parsifal at fourteen Bayreuth festivals from 1901, the year after Levi died, until 1930. On the other hand, Siegfried grew up with his doting father, ran the festival with his mother, Cosima (who was not only Richard Wagner’s widow, but also the daughter of Franz Liszt), and was a composer and conductor himself. In the years after World War I, it was Siegfried who traveled to New York to conduct concerts of Liszt, Wagner, and his own music to raise money for the Bayreuth Festival, which was always on the brink of financial disaster ever since its debut in 1876.
So, if you wish to prepare Parsifal as a young conductor, it would behoove you to listen to Karl Muck’s and Siegfried Wagner’s recordings from the late 1920s, when electric recordings had become the new standard. What you might expect from listening is not what you will find.
One’s natural expectation would be to hear a highly romantic, over-the-top interpretation, with swooping strings and full vibrato in the brass. Instead, you will find two distinctly different sounds. Siegfried’s is closer to a vaguely romantic ideal, with moderate tempos and the use of downward portamento, something that has been mostly banished from performance practice as being sentimental rather than purely expressive. Muck’s recording of extended excerpts from act 3 is far more idiosyncratic. Some of his tempos (the prelude, for example) are as much as twice as slow as are commonly heard. In a way, it is akin to what Stokowski would bring to this music, but without the exaggerated, vibrato-laden string sound Stokowski achieved.
The Berlin-Muck sound is slightly creepy—dead, one might say. The Berlin recordings made by HMV with Muck had many players who were students of the violin virtuoso Joseph Joachim, for whom Brahms composed his concerto. Joachim was the last violin virtuoso who did not use vibrato all the time, as is customary today. Muck’s Berlin Parsifal excerpts exhibit a cool sound that is significantly different from the Bayreuth recordings because many of the Berlin players were playing without vibrato much of the time.
This is the point: There is no unified, “correct” Wagner sound emanating from Wagner himself. Every city in Europe had its own unique traditions, and what was true in performing Wagner was true with every composer. During his lifetime Verdi in Paris would sound different from Verdi in Milan. Therefore, conductors should look at historical recordings as presenting us with options, parameters of choice.
The same pertains to two other recordings from this same period, though the work in question seems as far away from Parsifal as one can imagine. It is Stravinsky’s Rite of Spring. Once electric recordings made it possible to capture a symphony orchestra with some semblance of authenticity, HMV in France hired Pierre Monteux, who had conducted its world premiere in 1913, to record Stravinsky’s masterpiece in Paris in 1929. Stravinsky’s reaction was to record it himself, also in Paris, in 1929 for Columbia Records. We tend to give composers the benefit of the doubt when it comes to interpreting their own music, judging it to be definitive, but it is also important to ask a few questions before presenting them with that accolade.
When composers perform their music, they become the inventor and the salesman, and there is much going on inside them when they plead their case. If the composer is a weak conductor, it will always be difficult to decide if the results on a recording are what he intended, or merely what he was able to achieve. In the case of the 1929 Stravinsky recording, there is an inert aspect of his presentation and a pervasive lack of control that bespeak more of technical difficulties than of artistic intention.
Monteux shapes the melodies when that seems called for. Segments of the score sound like “normal” music—late Rimsky-Korsakov, perhaps. He occasionally adds dramatic accelerations toward the end of certain sections, something Stravinsky does not do. Was that because Stravinsky did not want to, or because he could not lead his orchestra into doing that? After all, Monteux had conducted all the first performances of Rite with the Ballets Russes, as well as its first concert performances. However, he was also conducting to Nijinsky’s choreography, which would have influenced his tempos, though we can never know how much.
Monteux was a master conductor, and if one listens to a much later recording (1951, with the Boston Symphony), it is astounding how much these two studio recordings are the same. The introduction in 1929 lasts 3:05 and the same music in 1951 lasts 3:03! His metronomic tempo for the section called “Augurs of Spring” in both cases is quarter note = 100. The cumulative difference in the total timings, separated as they are by twenty-two years, is twenty-one seconds. His muscle memory makes his recordings a Rosetta Stone of style and tempo of the music he premiered, and that includes many of the great orchestral works of the early twentieth century.
Stravinsky, on the other hand, seems a total amateur by comparison, though he does have an agenda. He clearly wanted to establish himself as the ultimate interpretive authority on the work that gave him his world fame, and who could blame him? However, when Stravinsky recorded it again in 1940 (to counter Stokowski’s version for the upcoming release of Fantasia), and then one last time in 1960 (when he had the opportunity to record his complete oeuvre in stereo), we hear three different Stravinskys. What they have in common is the composer’s desire to create the image that The Rite could always be contemporary: as tastes changed, so did his performance of the work.
The original narrative of the ballet score was very much a story and a description of ancient Russia. As that became less and less aesthetically acceptable, Stravinsky performed his Rite in a more and more detached way: brutal and mechanical. The most beautiful and seductive parts of the score were therefore submerged into an aesthetic that could not admit to it. In that sense Stravinsky’s three recordings tell us more about music history and its evolution than provide insight into The Rite. Since music has to be performed in order to exist, the evolution of orchestral playing, the composer’s evolution, and his improved skills as a conductor are all at play in this series of “authentic” recordings.
What happens when the composer is a great conductor? Here, too, the issue of the imprimatur of authority comes into question, along with how we might make use of the recorded documentation it affords. Richard Strauss, a great conductor, recorded extended excerpts from his 1911 Der Rosenkavalier while in London in 1926, where he conducted the British premiere of its silent-film version. His no-nonsense, straight-ahead approach to his own lush music would probably be viewed as insensitive today, whereas it is clearly the way the composer liked his music performed. Should this tempt a conductor to try this approach, or is it symptomatic of the idiosyncratic way Strauss conducted every piece of music?
No one exhibited this issue more than Leonard Bernstein; and the difference between what he asked of other conductors conducting his music and what he did when he performed it himself—and what he did when he recorded it in the studio—is a perfect case in point.
There is, obviously, no question of Bernstein’s abilities as a conductor. However, if we listen to the original cast albums of his Broadway shows, we can get a good idea of what he objectively wanted. Unlike his Broadway scores, he recorded all of his own symphonic works at the time of their composition, and therefore it would be difficult to differentiate between his recordings and the musical compositions themselves.
This was as much an artistic as a financial arrangement. His business was run by Harry Kraut, who managed to get Bernstein a payment for many things at once: the commissioning, the conducting fees, the television and recording fees, the royalties—and occasionally a separate fee for himself for producing the appearance of Leonard Bernstein. Because major conductors stayed away from conducting his music while he was alive, and since Bernstein was box office, having him conduct his own music was essential to encouraging audience interest.1
When I was music director of Candide in 1973, I asked Bernstein about Samuel Krachmalnick, who was the show’s original music director in 1956. Lenny said, “I remember how fat he was, how mean he was, and how good he was.” (Barbara Cook objected to this, saying that “Krachmalnick was never mean.”) Lehman Engel was the music director of the 1953 Wonderful Town, and Max Goberman served as music director of West Side Story. Maurice Peress was the original music director and conductor of Mass; and I conducted three different versions of Candide, the European and tenth-anniversary productions of Mass, and the three-act version of A Quiet Place, both its world premiere at La Scala and its American premiere at the Kennedy Center. In all cases, Bernstein was present. (He was not present for the recording of the sound tracks to the movie versions of On the Town or West Side Story.)
If, however, one compares the recorded evidence of Krachmalnick, Engel, Goberman, and me (Peress never recorded Mass) with Bernstein’s own recordings, one hears a completely different approach to these scores, and it is not a subtle difference. It is fundamental in Bernstein’s aching slowness, in its “explanation” of every twist and turn, in its reaching for epic importance. It always feels as if Lenny were making a case for his “light” music to be taken seriously and for his undeniable genius to be overtly demonstrated. It is as if there were a chip on his shoulder based on the years of criticism and denigration that fueled a certain defensiveness.
Mind you, if he had asked any of us to conduct his scores the way he would ultimately record them, we would have. When I want to hear West Side Story I will go to Goberman in 1957. Candide belongs to Krachmalnick. Bernstein’s own West Side Story recording comes from a time when he had become immensely serious, and his Candide was recorded when he was very ill, as was his cast. Most of that recording was made without the singers present. Bernstein laid down the orchestral tracks and the singers came into the studio later—and without his being present—to record their vocals. That it won a posthumous Grammy for him has more to do with the honor he deserved, rather than his representation in 1989 of a score he composed between 1956 and 1971. I love every minute of it, but I would never conduct it that way, even if I could.
Most opera lovers are aware that both Arturo Toscanini and Sir Thomas Beecham recorded La Bohème and that both recordings have profound historical roots and bona fides. Toscanini, after all, had conducted its world premiere in 1896, and Beecham had spent time with the composer, discussing issues of interpretation years after the premiere. These two recordings were made in New York City, Toscanini’s in 1946 and Beecham’s ten years later. Beecham clearly had something to say and wanted to confront the generally held view that Toscanini’s interpretation was the gold standard and that the Italian maestro “owned” La Bohème.
Indeed, these two readings are completely different in their approach and style. If anyone ever needed a lecture-demonstration on what makes one conductor’s interpretation different from another’s, this is it. As one would expect, Toscanini’s is significantly faster and tighter. Every tempo and quick change in the score feels welded to the others. It is breathless and relentless. Beecham breathes life and shape into every phrase. Who is right?
Timings of performances are a relatively simplistic way of showing similarities and differences among performances, because in reality what happens within every tempo is as important as the tempo itself. Emotional time is a very different matter from anything a clock tells us. Nevertheless, since La Bohème is a short opera in four acts, the running times of Toscanini vs. Beecham are illuminating:
Toscanini |
Beecham | |
Act 1 |
30:58 |
35:51 |
Act 2 |
16:58 |
18:12 |
Act 3 |
21:21 |
25:11 |
Act 4 |
24:42 |
29:00 |
These differences—a disparity of 15 percent—are not inconsiderable. Both men were in their late seventies when they recorded Bohème. Both were in a studio, not an opera house, with the singers near them. Toscanini was broadcasting live on the radio, which does of course make a difference since he had an audience in attendance. But it is the point of view that feels truly apart. If a critic’s review is autobiographical, it might be equally valid to view a performance by any conductor as also being autobiography.
In the winter of 1975, when I was rehearsing Fidelio at the Met, one of the house assistant conductors, the Viennese Walter Taussig, came up to me and said, “You know, my dear, there are many ways to conduct Fidelio, but really there is only one.” And with that he reached out with his right hand and, smiling, gently touched my chest.
Being part of the Metropolitan Opera’s history is a singular honor, especially when one gets to be broadcast from the house as part of its Saturday-matinee radio series. When our performances are broadcast and/or subsequently released, either by streaming or as commercial releases (discs, downloads, whatever), we are aware that what we are doing is being shared by a huge public—far larger than the number in the concert hall or opera house—and that the contents of these broadcasts will never go away. They are part of our legacy and endure the judgment of history.
What we do in a live performance, however, is determined by many factors that are irrelevant to the home listener. The most important of these is the power of silence. Erich Leinsdorf once wrote that silence is meaningless in a recording but an essential tool in a concert because of the effect it has on an audience. Every composer since Richard Wagner, and every conductor, knows one thing about silence: it is very powerful and must be used carefully. Sometimes it is a little bit of air between notes, called a Luftpause. Sometimes it’s even bigger: the “general pause,” indicated in a score by an empty bar with a G.P. on it. The longest silence of all is indicated by a symbol that looks like a half-circle floating above a single dot. This is known as a corona or a fermata. It is a stop of any length, to be determined by the conductor, who is empowered by the composer to decide how far to stretch nothingness, maintaining the music’s tension without losing the public’s attention.
Silence is the demarcation of sound. Once there is a silence, all that was heard before it is coalesced in the auditor’s mind. It also indicates that something (who knows what?) is about to happen. People occasionally refer to it as a pregnant pause. It is a technique used by political and religious speakers.
When Richard Wagner wrote the opening bars of Tristan und Isolde, he indicated that the tempo should be “slow and yearning”—langsam und schmachtend, as mentioned before. This is already extraordinary. Even though Wagner wrote very long operas, he hated slow tempos and rarely wrote the word “slow” in his music. That Tristan should begin this way makes it unusual. The audience hears a melody of only seven notes, the second of which is five slow beats long. Its length is only understood when the third note sounds. And when this mysterious figure concludes itself in a harmonic haze, one that makes its key and its function totally ambiguous, it is followed not just by a little silence, but seven beats’ worth. This silence is like a black hole for the audience, and when the second phrase begins, then and only then does the audience have a sense of the immense size of the temporal universe that is Tristan.
Like so much that Wagner achieved, the influence of this bold opening gesture affected just about everyone who composed after him. Debussy repeated the idea in his Prelude to the Afternoon of a Faun, when, after its famous flute solo and the response of the orchestra, there is a total of six “very moderate” beats of nothing. The brain is alerted to an input change. The ear, searching for more information, opens up to perceive what is next. Silence brings an audience together—and sometimes does it shockingly so.
Stravinsky ended part 1 of The Rite of Spring, which is as loud as anything in unamplified music, with an abrupt stop. At its premiere performances in 1913, the last note was synchronized with a freeze from all the dancers, hands in the air and in a tight circle. Lights out. Curtain. After a few moments, conductor Pierre Monteux began the orchestral introduction to part 2. But when Rite is performed as a concert piece, there is no way to prepare the audience for the sudden silence. It is like running wildly and suddenly falling (in slow motion) off a cliff.
Theatrically savvy composers are always writing, or expecting, silences in their music. Verdi’s overture to La forza del destino is a perfect example: Three blasts from the brass are followed by silence; three more blasts, and silence. The conductor determines the length of the silences by what he feels from the audience. (Verdi put a fermata on the bars of silence.) If the audience is not yet settled, one waits and waits. After the second set of three blasts, the audience understands. Nevertheless, we wait, because what follows is the whispering theme of the perilous journey of the heroine, Leonora, and it needs total collective concentration of the public to hear and to listen. The nervous quietude of this figure makes the crescendo and its ultimate goal of fortissimo absolutely shattering, provided it starts practically inaudibly. Without an audience, none of this makes much difference. Verdi knew the performance and the work itself depended on the participation of a live public. A recording of it simply cannot make much use of the silences. They are inert and not kinetic. With no sense of danger, a recording is experienced personally and not societally.
Recordings are a great equalizer of sounds and densities. A string quartet can be as loud as a hundred-piece orchestra. That is due to the mastering of the album—the process of editing the various takes, mixing the various tracks to create a good balance, and adjusting the volume so that it is neither too soft to be easily heard nor so loud that it distorts. Of course, you can simply adjust the volume control yourself. When my Music for Alfred Hitchcock was released in 2015, based on two live performances with the Danish National Symphony Orchestra, we left the volume at a consistently realistic level for the entire disc. That meant that the loud moments in “Psycho—A Narrative for String Orchestra” were quieter than the loud moments in “The Storm Clouds—Cantata,” which is scored for a full orchestra, a mezzo-soprano soloist, and chorus. The audience in the live performances makes the adjustment of accepting two entirely different objective parameters of “loud” and “soft.” In a recording, one expects all louds and softs to be of equal decibel parameters. Perhaps mastering the album this way was an error, though it is a far more honest document than pumping up the volume whenever it fell below the ranges of the works that surrounded it.
Even as the technical advances of recordings have improved vastly ever since the analog era morphed into the electric era and then the digital era, one thing is never present in a recording: a true representation of the density shifts in orchestral music. Gustav Mahler was a revolutionary in his use of chamber-music sections within the epic sounds of his symphonies, but we never hear that in recordings. The engineer will inevitably bring up the microphones on the solo instruments. If he does not, you will do that yourself—especially when listening on headphones or driving in your car.
Beethoven, for example, made brilliant use of density shifts in his Symphony no. 5. If I asked you who plays the opening notes of Beethoven’s Fifth, most people would answer, “The entire orchestra.” In fact, it is just the string section and two clarinets on that famous dah-dah-dah-DAH. Beethoven delineates what “very loud” means with those famous four notes, marked fortissimo. Having established fortissimo for the listener, he follows it with the other end of the spectrum: the second violins then play piano. The contrast is dramatic. After the strings call and answer the famous rhythmic figure, a brief and violent crescendo occurs in one measure when, for the first time, the flutes, the oboes, the horns, the trumpets, and the timpani enter, going from piano to forte in four short notes, followed by the opening motif. This time the fortissimo involves all the woodwinds and horns with the strings. Therefore, that forte Beethoven has just given us will be louder than the fortissimo that started the symphony, and this second fortissimo will be louder than everything that has preceded it.
No recording captures that. The opening four-note fortissimo on a record will inevitably be as loud as the second fortissimo, even though the second one uses more instruments. If the engineer did not do that, the passive listener in his or her living room or wearing headphones in the subway would find the music lacking in energy. The dynamic ranges are set at unnatural levels because recordings deliver their information in a very different way from live performance.
The greatest invention in Beethoven’s attempt to “burst our breasts with full-voiced harmonies,” in the words of the first review of the symphony, written in 1810 by the author and composer E. T. A. Hoffmann, is the arrival of the fourth movement out of the third. Beethoven famously does not compose a break between these two movements. Instead, a mysterious figure, marked sempre pp, remains static and ambiguous, as if the composer were unsure of his goal, or perhaps were playing with us. We are firmly in C minor, the key of that famous first movement, and Beethoven delays and delays until, in less than eight seconds, the music grows in loudness to a fortissimo that is from an entirely different and expanded universe of sound.
Not only are we in C major, the key of “joy and sunshine,” as Beethoven once said, it is a C major that includes a first entrance of a piccolo on the top of the chord, expanding the woodwind harmonies an octave higher than in the previous movements; a first entrance of a contrabassoon, which expands the woodwinds an octave lower than anything they have played up to this point; and three trombones, which have been silent for twenty-five minutes. Suddenly and unexpectedly there is an increase in the density of the harmonies and the overall power of the fortissimo to an unprecedented level. The floor drops, the ceiling rises, and we are in an entirely different space: a grand, sun-filled room never imagined until this moment.
Again, no recording has ever captured this. If the entire symphony is engineered to make the opening of the last movement what it is in concert, you inevitably will have turned up the volume earlier on, and once movement 4 begins, you will have already experienced sensory overload—or someone you live with will be yelling, “Turn that thing down!”
That said, once we accept that a studio recording is (1) not a performance and (2) not a piece of music, something new comes into play in the recording studio that replaces the danger and the electricity of the shared live experience: total control of the document—its shape, its sound, and its text. You no longer have to depend on inconsistent acoustics or random technical errors. It is critic-proof. And it lasts forever.
The recording studio, the film-scoring stage, or the auditorium that is fitted out with microphones, closed-circuit television, headphones, and telephones are as magical as any room. The morning of a recording is filled with electricity and anticipation: The participants are going to inscribe music for history. Everyone knows why they are there, and everyone feels it. When the red light goes on in the room and the engineer announces “Take 1” over a loudspeaker, the room becomes profoundly silent. No one moves a chair or coughs. There is no audience. The conductor waits a moment, and the concentration in the room is palpable. Then there is music.
The engineer and the producer are behind a glass: one is following the score and taking notes on anything that seems wrong, and the other, also following a score, is listening carefully to ensure that all the parts of the orchestra are being picked up on the various tracks that are recording the musicians. Is the triangle there? Are the harps present? How much of the room’s sound is being recorded? Do the percussion sounds leak into the woodwinds, or are they discrete enough to balance them properly at a later time? Was there distortion in any of the tracks?
At the first break, the orchestra members go their separate ways and the conductor enters the recording booth, sits at the console with his score, and listens to what has just transpired. There is never enough time to listen to everything and get back to the session, since the orchestra breaks are shorter than the time it will take to play everything back and discuss it. There is no break whatsoever for the conductor and those running the session, who will have to make any desired adjustments on the fly. The first takes are usually discarded.
And so it starts again—which may mean going back to something that was not “covered” in the previous takes, or going forward, with the producer keeping a careful eye on the clock and how much music must be recorded in the allotted time. Privately, the producer could be in a panic. “God, that was awful!” he might mutter with the in-house microphone turned off. Then he might announce on the speaker system, “That was terrific! Let’s try it once again just for safety, can we?”
When the Hollywood Bowl Orchestra and I had finished what we all thought was a magnificent take of Miklós Rózsa’s waltz from Madame Bovary, and the composer had just said, “Perfect!,” I got on the phone to the recording booth and our producer, Michael Gore, said, “ ‘Perfect’? That’s fine, but we will need to do it again.”
As the sessions proceed, a composite “performance” emerges. Inevitably, there is one last run-through as time is about to run out, and it becomes the basis of the recording. Of course, when the work being recorded is a concerto, the soloist is a part of every playback session. When it is an opera, as one would expect, the process is enormous.
Albums are not recorded in the order you hear them in a final commercial release, so when you read that a soprano “takes a while to warm up” on a recital album, recorded in a studio, the critic simply does not understand how records are made. Operas are never recorded in order, either. Sometimes even a single aria is not recorded in order! In a number of cases, singers ask to record the climactic ending first to ensure that their voices are fresh. After all, they are singing more in the studio sessions than they would onstage. This is their legacy, too. Record companies that regularly recorded famous opera singers sometimes had a singer come in and record an entire series of high Cs—or notes even higher—sung on different vowels, and put them in their audio bank just in case they needed to slip them in some time in the future.
When a conductor records a work he has never performed before, he has to imagine its long arc and fit the segment being recorded into that imaginary journey. This, perhaps, is the most exciting and creative aspect of studio work for a conductor. Georg Solti apparently had never conducted Siegfried before he made the definitive recording on Decca with Nilsson, Wolfgang Windgassen, and Hans Hotter. Needless to say, the entire Decca series of Entartete Musik—the music banned by the Nazis—was being played, conducted, produced, and recorded for the first time.
That summons forth another kind of magic: playing music that has been dormant for generations but suddenly comes to life as an orchestra reads through it for the very first time. There is no more moving moment for a conductor than experiencing the myriad of invisible and unique sonic patterns in the air that have not existed for sometimes more than a half century, and the feeling of everyone putting his or her life experience and talents into awakening Sleeping Beauty. In the case of the music banned by Hitler and the personal stories attached to each work of art, and the human cost to the composers and their families, it can be overwhelming.
When I struck up the opening bars of Korngold’s Between Two Worlds with Berlin’s Rundfunk-Sinfonieorchester (in what had recently been East Germany but was now part of a reunified Germany), the orchestra had never played a note of the composer’s music. They were reading from photocopies of the parts that had been played only once before: by the Warner Bros. Orchestra in 1944 when it was recorded under the composer’s direction for a film about people on a mysterious ship fleeing the Blitz in London, who do not realize they are dead and awaiting judgment. These superb German musicians had all been trained in a city that had been the capital of the Third Reich when this music was composed and had all been trained when that same city was the capital of a Communist German country where classic American films were rarely seen. They were now playing this music, and Korngold was somehow alive again and free. Note after note, harmony after harmony—a cymbal crash, a timpani roll, an aching melody—music unheard in Germany filled a room, and somehow the world seemed a better place than it had been a few moments before.
Once the recording sessions have ended, the editor assembles the best takes into a rough edit. The conductor, producer, and engineer listen to it, send in corrections, and back it goes to be mixed. Mixing is the most fun, because it is conducting one’s own conducting. Here is where Karajan and Stokowski put the finishing touches on their work. Bernstein depended on John McClure to create the “New York Philharmonic sound” we all recognize—brilliant, a little metallic, and in your face. Karajan’s “sound” was not just the Berlin Philharmonic in the Jesus-Christus-Kirche for Deutsche Gramophon (the same venue in which Decca’s Berlin recordings were made). It was a combination of his DG engineers’ knowledge and creativity (where they placed the microphones, and what kind of microphones were being used), the acoustics of the room, the basic sound of the orchestra, and the overall aesthetic of the conductor. In Karajan’s case, it was not as specific in pinpointing individual instruments, as in Bernstein’s recordings; rather, it was amorphous and darkly glowing.
The famous Decca engineers partnered with Solti and producer John Culshaw for the uniquely visceral power of their Vienna Ring cycle. Like performing music, recording it in a studio is about partnership: the musicians, the conductor, the producer, and the engineers.
The room in which the recording is done is also important. In Hollywood, I chose its greatest room for recording the Hollywood Bowl Orchestra: the room built by MGM when film scoring first began, a room that had not been touched since the 1930s except to update the equipment. This was the room in which Judy Garland sang “Over the Rainbow” and Miklós Rózsa recorded Ben-Hur. Our engineer, Joel Moss, used rebuilt analog microphones for our digital recording to bring a classic warmth to the sound, and producers Michael Gore and Tommy Krasker shared a collective vision for the final product: The Hollywood Bowl Orchestra recordings on Philips were to place the listener where I stood, front and center. It would be a combination of the natural warmth of the MGM scoring stage (now owned by Sony) and the specificity of illuminating the orchestration through subtle adjustments of the levels from individual microphones.
Because my colleagues and I felt that these studio recordings were an art form that did not intend to replicate live performance, we were able to mix the already excellent playing of the orchestra so that the listener would hear what I was hearing and imagining. Every murmuring thirty-second note in the woodwinds at the opening of “Dawn” from Daphnis et Chloé is precisely present, with the harps separated left and right. The chorus, marked “offstage” by Ravel, begins in the distance, and we slowly brought their sound into the room until the last, climactic one, in which they are totally present before you, the listener.
With another dawn—the finale from Schoenberg’s Gurrelieder—we recorded the gigantic orchestra by itself. Then we brought in the Los Angeles Master Chorale, using only a small group of superb singers rather than the more than 150 voices required for this work. (When Gurrelieder is performed live, there will be an orchestra of 150 on the stage as well as the equally large chorus somewhere in the hall. Recordings of the work maintain this performance setup, and inevitably the chorus is late and flat. Mixing it is all but impossible, since the venue for the recording will most likely be a cathedral or a town hall with a great echo and vast distances between the conductor and the singers.) With a headphone on one ear (to hear the playback of the orchestra), and the other free to hear each other, the singers performed their parts, conducted by me and synchronized to my own recording of the orchestra. And then we recorded the chorus again over their previous take—making sixty singers sound like 120—and then did it a third time. The result for me is a recording of complete clarity and epic proportions. It is a studio recording that creates an aural replication of what Schoenberg put on the page, and (I believe) imagined.
One of the great challenges with minimalist music is that when it is in a groove it has to be perfect, chugging along with its meters changing all the time and its surprise rhythmic bursts, but always sitting on top of something inflexible and almost impossible to consistently achieve due to its machinelike demands. Much as we might love this music, performing it is more like following a strict click track in a movie soundtrack, but without the click. Any false entrance or variant of the tempo is an obvious mistake, and the fundamental energy of the music dissipates.
Again, we used technology to solve that challenge with John Adams’s The Chairman Dances. This kaleidoscopic work has two kinds of tempo challenges: the strictly inflexible pulse sections and the contrasting passages full of the ebb and flow of romanticism. Although the orchestra had played the work for its Bowl premiere, once we were in the studio the most efficient way to record it was to use clicks for those sections that held to a single tempo, record the rubato sections as we would normally record, and then edit the two together. It is a studio recording and is a document.
Thus, recordings can be an art form all their own. The source is always the playing of instrumentalists and the singing of singers. It can be an artifact. It can be arti-fiction, but really, if you are creating something that is not ephemeral, do you want to hear the soprano go flat every time you hear her sing “Un bel dì”? No. You want it to be edited. No recording of Renata Tebaldi will ever replace the memory of her singing that aria at the old Metropolitan Opera House in 1960. When she arrived at the climactic high B-flat, the fullness and warmth of her voice created a resonance in the auditorium as every molecule of air vibrated in sympathy with her personification of futile hope. That great room, with its history and embracing acoustics, loved her as much as we did in the audience.
That performance is mine, stored somewhere in my brain, and you will never hear it. Together, however, we can share in her recordings of that aria, played through speakers on our walls or strapped to our ears.
When television was young, it seemed as if it would be the porthole through which all things could be seen and heard: cooking shows, puppet shows, farming shows, old people talking (Life Begins at Eighty), history shows (You Are There), news (for a total of fifteen minutes), symphonies (NBC had Toscanini’s orchestra of course), operas (NBC had an opera company!), dramatic shows (Playhouse 90), variety shows (Ed Sullivan’s Toast of the Town), and sports (including female roller derby). Somewhere during that time, the president of the American Federation of Musicians, James Caesar Petrillo of Chicago, worked hard to ensure that professional musicians were being paid better and were being protected from many of the unfair practices in the preunion days of music (meaning all of history prior to 1940).
He also stopped all commercial recording in America for a number of years in order to get higher wages. The work, instead, went to Europe. He also famously said that the Chicago Symphony would never be on television. What he and his supporters believed, and many people still do, is that the way to ensure an income for professional musicians is to limit their accessibility in recording and broadcasts (“Support Live Music!” was a motto), thereby preserving their value in live concerts. The union also did not want nonunion amateur ensembles to infringe on the territory of the professional musician. In general, broadcasts, and above all recordings, were seen as costing jobs.
Sports chose the opposite path. Its leaders wanted kids to play early and keep on playing, with their parents as cheerleaders. Amateur sports became a huge business with a gigantic audience on television. In America, not only does amateur football feed professional football, but billions of dollars are currently made from people who are happy to put down their money to support the entire gamut from peewee football to the Hall of Fame.
No one and no technological advance will ever get rid of the live concert or a live athletic event. Humans are social, and we rather like—need—to gather together. Provide people with the incentive to experience something together, and they will. On the other hand, as Sam Goldwyn once wisely said, “If people don’t want to come, there’s no way to stop them.”
While a number of artists, and the music unions, object to audience members (fans) taking out their personal electronic devices to record parts of a performance, especially when it is a distraction for the artists or the public, we could also view it as both moving and uplifting. Each person wants to hold on to something magical and dazzling—fireflies in a bottle. It is not for the purpose of financial gain or stealing something. It is about preserving some part of the magic they are experiencing—specifically from their point of view, their place in the hall, with all its attendant room noise, the cheering, and the blurred visuals. Those recordings on personal electronic devices are like trying to “hold a moonbeam in your hand,” as Oscar Hammerstein II once said. It is a childlike hope of permanence, and recordings provide an imperfect pathway toward that impossible dream: to remember, to recollect, and to commemorate.
All music lovers listen to recordings. It has changed everything involving how we listen to music because it affords endless accessibility and the possibility to hear what we love again and again. Somewhat like the movies, it provides a platform for sharing art with millions of people without the limitations of time or place. But it has also created certain performance traditions, many of which emanated from the late 1920s and early 1930s, when electric technology began the modern era of recording. These traditions have been transferred into an audience’s impression that the recording is the piece of music, which it is not.
It has also created expectations of technical perfection that are impossible to fulfill in performance but are usually mitigated by the excitement of an audience sharing the unique experience of the live concert in an event that cannot be recreated ever again. When Joan Sutherland made her Met debut in Lucia di Lammermoor in 1961, the sold-out audience gave her a huge ovation upon first seeing her onstage. She stopped and waited, and waited…and waited as the ovation built and ultimately subsided. Subsequently, she said she was terrified that the living Joan Sutherland would not measure up to the Joan Sutherland the New York public knew only from recordings. That night she was even better, precisely because 3,850 people watched, listened, and participated. She was theirs and they were hers for three hours that are uniquely preserved in the memories of those who attended and survive.
If he is so blessed to become a conductor, a boy in the audience at that debut can grow up and honor Joan Sutherland forty-three years later—and say “Thank you.”
Dame Joan Sutherland, me, and Marilyn Horne, at the Kennedy Center Honors, 2004