The Fifth Element #7

Teresa of Ávila (1515-1582) is acknowledged as a saint by the Roman Catholic Church. She is additionally accorded the rare (especially for a mystic) distinction of recognition as a "Doctor" of the Faith. On a somewhat less exalted level, but perhaps resonating even more clearly with the truth of common human experience, Teresa (who had Jewish ancestry; why is that not surprising?) is credited with coining the phrase "Be careful what you pray for, you might get it."

Even before the Wilson Benesch Discovery loudspeakers I wrote about in my January column had cleared customs, Teresa of Ávila had in some cosmic way become patroness of this project, which, if your memory needs refreshing, was to put together an extremely high-resolution system capable of making the most revealing comparison possible between the SACD remastering of Glenn Gould's 1981 Goldberg Variations (itself a re-recording of repertory Gould had blazed a trail with in 1955) and its "Red Book" 16-bit/44.1kHz CD original.

Be careful what you pray for...
Gould's 1981 performance was the result of a protracted and thoroughgoing reconsideration of the aesthetic premises of the legendary recording of the Goldbergs he had made in 1955. To simplify to a criminal extent (in view of the sprawling nature of the subject matter) (footnote 1), Gould came to believe that his 1955 recording failed—through a "lack of deliberation"—to do justice to what he increasingly perceived as the mathematical relationships creating an essential unity among the variations.

Once having posited a common numerical basis for each section's structure, Gould concluded that the contrapuntal variations, in particular, needed the breathing room that a slower, steadier musical pulse can provide. Each listener will have to decide for him- or herself (with the proviso that the more studying you do, the more likely your conclusion will be worthwhile) whether the 1981 results are the fruit of artistic growth, or the product of unnecessary, neurotic self-criticism.

Regardless, the 1981 recording remains uniquely fascinating and evocative, not the least because of Gould's humming and singing. (And also because of the poignancy of Gould's having died of a stroke before the recording could be released.) I have mentioned before that one of my touchstone equipment-evaluation tracks is the Aria of Gould's 1981 Goldbergs on CD, in large part because Gould's humming and singing challenge both a system's resolving power and its timbral accuracy. Gould himself was as troubled by his humming as he was by his inability to curb it, in which effort he even underwent psychotherapy.

Pluses and Minuses
My initial impressions of the SACD remastering of the Goldberg Variations were paradoxical, to say the least. On the good side of the ledger: the exceptionally dynamic, percussive start of Variation 1—in the past—has always splintered into pieces, to the extent that I had concluded that this was due to uncorrectable digital overload. (Gould really nails it, to emphasize that Bach was not following the usual course of taking "baby steps" while building variations of increasing complexity and range.) On the SACD, Variation 1's punched-out first note (which I still think jumps the bar line a bit) is now just a questionable artistic decision and not an engineering goof. So far, so good.

The minus side of the ledger is where the paradox manifests itself. The SACD has more resolving power. Unquestionably. One visitor said it sounded like Gould's humming had been "turned up in the mix." I certainly hope that is an impossibility!

But there is also more noise. The noise level so surprised me that I phoned John Atkinson to tell him about it and discuss possible explanations. When I said that the SACD had a lot more noise, John (giving me more credit for such things than I deserve) initially thought I was looking at a spectrum analyzer, and commented on the substantial out-of-bandwidth (ultrasonic) noise that is a signature of DSD's sampling rate and filtering. Feeling a mite sheepish, I said, "No, through the speakers." Ah-ha.

Regrouping a bit, JA speculated that, given the age of the recording, had it been done by CBS on an early Sony digital system without an S/PDIF output, Sony of today might not have been able to get a digital signal directly off the video tape. But that would mean only that the converter's analog outputs were used—which should not make any more noise than any other piece of professional audio equipment. What I was hearing was more like analog tape hiss.

That opened up the possibility that, as was the case with many early digital recordings, editing equipment was not available, or not up to certain tasks, so the signal may have been bounced to analog for certain edits, then put back to digital. Were that the case, one could then speculate that the original CD release may have involved some filtering to erase those fingerprints, so to speak (see Sidebar).

I also must confess that a truly rigorous comparison between "Red Book" and "Red Book redone for SACD" must wait for another day, because the Accuphase DP-100/DC-101 two-box SACD player I was using for these comparisons employs a non-defeatable upconverting scheme when processing "Red Book" data. Drop in the CD and what you hear is the CD, upconverted in the PCM domain. Ah me. Nonetheless, differences remain. The SACD is more detailed, while the CD—upconverted—is a touch warmer.

What threw matters into a bit of a tizzy was my reading (and participating in) a thread on the Audio Asylum's SACD bulletin board about a "ghost orchestra" on the 1981 Goldbergs SACD. I had already observed that the noise seemed somewhat variable, and even seemed to "pump" on some tracks, but I had never listened loud enough to make out (Good Lord, how loud those people must listen!) what is, on close examination, yessirree, an orchestra playing in the background of track 16: Variation 15. In the right channel only. Now, when I say in the background, I mean background. As in, Gould is playing the piano in your room just about as loudly as a piano can be played, and three rooms away, someone is quietly listening to classical music on a tinny radio, with the door almost closed.

Footnote 1: The simple, wistful Aria that is the subject of the 30 variations is most likely not even by Bach. Bach was no mere tunesmith; his creativity here consists in crafting variations not so much on the melody but on its underlying harmonic structure, especially the movement of the bass line. Bach's genius lay in recognizing that hidden potential, and also in recognizing that, after stretching that potential to include everything from virtuosic exaltation to searing tragedy, the human spirit experiencing this work would need the anchoring consolation of a reprise of the Aria in unadorned form. There is a classy website dedicated to the Goldbergs.