Treasures for the taking
SHOULD SONY HAVE RELEASED GLENN GOULD’S OUTTAKES?
Sitting comfortably in my living room, I had every intention of listening to the boxed set that had recently landed with a thud in my mailbox. For Bach lovers, Glenn Gould fans and pianists (I’m all three), here was perhaps the ultimate find of recording history, finally available: the nearly five hours of Gould’s complete recording sessions for his momentous debut album, Bach’s Goldberg Variations.
But I couldn’ t make it through, not even close. Let me explain.
In four intense sessions in a Manhattan studio in June 1955, the 22- year- old Gould recorded his breathless, uncannily clear account of the Goldbergs. The release turned what had previously been considered a lengthy piece for harpsichord, of interest only to Bach specialists, into a runaway hit. It also made the gangly, eccentric Gould an unlikely classical superstar.
The tapes of those sessions were stored for years in the archives of Columbia Records, now Sony Classical. But in a 2004 interview, the recording’s producer, Howard H. Scott, said it would be a “disgrace” to make them public.
“If Glenn knew Sony Classical was going to release those outtakes, which he rejected,” Scott said, he “would probably come down and shoot anybody who allowed them to be released.”
Well, Scott died in 2012. And Sony Classical has gone and done it. The bulky set, issued in September, includes five discs with all of Gould’s takes, a thick book of photos, documents, articles ( and that 2004 interview with Scott), and an 80- page guide listing every single take, interspersed with the transcribed spoken exchanges between Gould and Scott.
Though Gould might have objected, I have no problem with Sony’s decision to release these sessions. ( A few selected outtakes were actually included in two previous Gould Goldberg reissues.) In an essay, Robert Russ, producer of the new set, argues for the documentary value of the tapes, which he deems part of our cultural legacy.
Fair enough. But who is the target audience for this? When I first received the set, I wondered whether even the most devoted Bach and Gould lovers would actually listen to all 282 tracks. The final recording is exhausting enough: Ignoring all the structural repeats in the score, Gould brought the Goldbergs in at just over 38 minutes, while performances that adopt more conventional tempos and observe all the repeats can last up to 90.
It’s still amazing to hear the seemingly impossible clarity of Gould’s playing, the sometimes manically fast tempos. And, for all his frenetic energy, in passage after passage, he brings out the music’s majesty, dancing grace and tenderness. Hearing the intense young Gould at work during these arduous recording sessions, playing through a variation at a breakneck tempo with prickly sound, then playing it again, and again, and again, is not just exhausting; it’s stupefying. What, I asked myself, was the point?
And yet the more I dipped into these takes, the more revealing they seemed — not just of Gould’s process, but also of the conflicting imperatives of making a recording. Master pianist Artur Schnabel, for one, initially resisted entreaties to make records. The nature of a performance, he wrote in a memoir, “is to happen but once, to be absolutely ephemeral and unrepeatable.”
Schnabel eventually decided that the potential benefits outweighed the downsides. His discography includes a probing survey of Beethoven’s 32 piano sonatas, recorded in London during the early 1930s. Listening today, you might wonder why he approved some of the takes. The demonic, perpetual-motion finale of the Appassionata Sonata, for example, has stretches of hectic, rushed and smudgy playing.
But Schnabel’s account also has exciting sweep and wondrous shadings. The f l aws wouldn’t have bothered him a bit in a live concert. Clearly, in this recording, he was trying to main- tain a semblance of that “ephemeral and unrepeatable” immediacy. Perfection wasn’t the point.
The artists who are usually the easiest to record are those who report to the studio with a thought- through and fully prepared performance. All they need to do is repeat it for the microphones. This may explain why Arthur Rubinstein, who simply loved playing the piano for people, also felt at home in a recording studio. Over his long career, Rubinstein matured as an artist and altered his thinking about many of the works he played. But in the moment of making a recording, from all reports, he was content simply to lay down his current concept of a piece. That was that.
That was not Gould. As Scott explains in that 2004 interview, Gould “changed a lot as he recorded because he wanted to try different tempos, different accents, different phrasings, because that’s why he loved recording so much.” His restless, searching mind is part of what made his playing so engrossing and original. But it’s also what made it challenging, Scott said, to complete the Goldberg recording.
Gould’s restlessness courses through every track ( at least those I listened to). He begins by playing the opening aria at a gently lilting tempo and with pristine clarity, though a hint of intensity hovers within.
“Listen, hey, who the heck tuned this piano this morning, listen to this,” Gould says, thumping on a note that sounds off.
In Take 2, he plays the aria more fleetly and bright ly, though the tempo slows a little as the music unfolds. Still, he shaves eight seconds off the timing of Take 1 (which was two minutes seven seconds). In Take 3 he teases out certain crucial notes in the left hand; in Take 4, to me, the melodic line sings out more prominently. In Take 5, he plays with a noticeably lighter touch and thinner sound.
“I think that was it,” he says, “but ….”
It was not it, as we discover. There’s a sixth take that day, and then 11 more during the final sessions, for the repeat of the aria.
The first variation, which sounds jaunty in most performances, bursts forth in Gould’s first take with stunning speed and incisive articulation. “I think that’s it,” Gould tells Scott. “But we might try another one.”
The producer says that it was beautiful, but ventures, “Only I must tell you ….”
Gould guesses the problem. “You mean the voice?” he asks. ( Gould famously could not prevent himself from singing along when he played. And his voice, though muffled, comes through now and then.)
He tries again. But now he’s clearly rattled, as you can hear in some edgy notes and general uptightness. “No, now you’ve got me a mental block, same thing happened,” he says.
Take 3, played at Gould’s desired racing tempo but now with brio and command, would be the approved one of the first variation. Still, Gould plays more takes — 13 in all! — between which he often expresses exasperation with himself.
As I kept listening to take after take, my mind froze. After a while, I stopped being able to figure out why Gould and Scott finally went with certain versions and not others. The entire recording could have been edited much differently and still created a sensation. And, in fact, Gould grew to feel that his 1955 Goldberg recording was, he said in an interview near the end of his life, “just too fast for comfort.” In 1981, he had recorded the work again, with a still bracing but generally mellower approach. Overall, this is the version I prefer; it was released the following year, days before he died of a stroke, at 50.
Over his career, Gould increasingly embraced recording’s potential to foster experimentation. He gave up playing concerts in 1964 and retreated to the studio, where he got involved with the detailed engineering of his releases, sometimes juxtaposing different portions of a piece played with various styles and approaches into a curious final synthesis. In a 1966 article, The Prospects of Recording, he fantasized about a future when listeners would be granted tape- edit options and could patch together their preferred versions of, say, Beethoven’s Fifth Symphony, from recordings by different conductors.
A happy compromise has emerged over the past couple of decades, as sophisticated digital technology has made live recordings, which are cheaper to produce, almost preferable to studio ones. Engineers can record an orchestra in a series of performances of, say, an epic Mahler symphony, and then select one version, essentially, for release. But through digital editing, they can fix any problems, down to replacing a single wrong note with a correct one from another performance. The best results combine “ephemeral” immediacy with good sound and perfect (or perfect enough) execution.
For me, the most poignant parts of the Goldberg sessions come when he stops for a moment to practise a passage — sometimes slowly, sometimes with just one hand. Here, for an instant, he seems like all dedicated pianists, trying hard to bring clarity, a particular touch, lyrical shape or rhythmic bite, to a phrase. You get a fleeting, revelatory glimpse into Gould’s feeling his way into this music.