The Recording Angel

Back in the last century, I mused in this space about the essential difference between recorded sound and the real thing. I had been walking to dinner with friends when I heard the unmistakable sound of live music coming from a window. But here was the kicker: rather than the instruments being of the audiophile-approved acoustic variety, they were two amplified electric guitars. Their sounds were being reproduced by loudspeakers, yet it was unambiguously obvious that it was not a recording being played through those loudspeakers, but real instruments.

"What on earth can be the readily identifiable difference," I wrote in 1995, "between the sound of a loudspeaker producing the live sound of an electric guitar and that same loudspeaker reproducing the recorded sound of an electric guitar?" I went on to conjecture that the act of recording inevitably diminishes the dynamic range of the real thing. The in-band phase shift from the inevitable cascade of high-pass filters that the signal encounters on its passage from recording microphone to playback loudspeakers smears the transients that, live, the listener perceives in all their spiky glory. And as a high-pass filter is never encountered with live acoustic music, that's where the essential difference must lie, I concluded, quoting Kalman Rubinson (who had not yet joined the magazine's team of reviewers) that "Something in Nature abhors a capacitor."

But two more recent experiences suggest that there must be more to the difference than the presence of unnatural high-pass filters. At the beginning of November 2009, I drove down to the Maryland headquarters of cable manufacturer Stealth Audio to take part in two evenings' worth of live vs recorded music. The idea was simple: two pianists, Genady Zagor (pictured in Wes Bender's photo) and George Vatchnadze, would perform for an hour on a well-prepared Steinway D grand piano, which I would record. Then, after a break, the audience would hear the recital again, this time played back through a system comprising Vivid Giya G1 speakers driven by Convergent Audio Technology JL3 150W tube monoblocks, a CAT SL1 Ultimate preamp, and Esoteric's five-box digital front end. Cables would all be the ultra-expensive, helium-filled Stealths.

And to ensure the highest possible digital quality, both the dual-mono Esoteric X-01VU D/A converters and my two two-channel dCS A/D converters, running at 24 bits resolution and a 96kHz sample rate, would be locked to the ultra-high-precision Esoteric G-0Rb "atomic" master clock. The digital data would be converted to FireWire with Metric Halo MIO2882 and ULN-2 interfaces and, using the Metric Halo Record Panel software, stored as four mono AIF files on my Apple MacBook.

My microphones were a spaced pair of high-voltage DPA 4003 omnis and an ORTF pair of DPA 4011 cardioids, all four amplified by low-noise Millennia Media preamps. I set up the mikes closer to the piano than I would normally choose, because playing back a recording in the same room in which it was made results in a double hit of that room's acoustic. This can introduce an obvious difference between the live and recorded sounds that will work against the illusion. And, of course, arranging for a room that would allow the piano enough reverberation to "breathe," but not so much as to hobble the loudspeakers, was not trivial. Stealth's Serguei Timachev had done a great job on the acoustics of the room, however, while Philip O'Hanlon, of Vivid importer On a Higher Note, worried away at the positioning of the Giyas until, by the second night at least, he felt they were working with rather than against the room.

The recital program covered a variety of pianistic styles. George Vatchnadze opened with two Scarlatti sonatas, followed by Chopin's fourth Ballade in f, and Scarbo from Ravel's Gaspard de la nuit. The closer was Genady Zagor performing Mussorgsky's Pictures at an Exhibition. The live sound of the big piano in the fairly small room, which seated 20 or so listeners, ranged from intimate in the Scarlatti to viscerally explosive in the Mussorgsky. At the climax of Pictures, The Great Gate of Kiev, with Zagor pounding the keys, the walls literally shook. But the Esoteric-CAT-Vivid-Stealth system got right not only the tone colors of the real thing, but, to my surprise, also both the loudness and the microdynamics. The transients were sufficiently spiky, and the impact of the recorded piano was as viscerally overwhelming, as the real things had been. Certainly the audience seemed impressed by what they heard.

However, as pleased as I was with both the recording and the playback system, the "bigness" of the 9' Steinway had been diminished. Even with the same tone colors and sound-pressure level, the instrument and the loudspeakers were exciting the room very differently.

The second experience was hearing the Yamaha AvantGrand N3 piano, which Bob Reina enthuses about in this issue's "Industry Update" (p.16). Whether it was listening to Bob improvising, or to a piano-roll transcription of Rachmaninoff playing his arrangement of Fritz Kreisler's Liebesfreud, the sound of this hybrid electric/acoustic instrument, in which loudspeakers replace the strings, was very convincing—at least, until the same piano-roll transcription was played on a 9' Yamaha concert grand fitted with a Disklavier mechanism. Again, while the electric instrument got both tone colors and loudness correct, it didn't sound as physically large as the real thing, and its relationship with the room acoustic was still different; the "bigness" was closer than it had been in Maryland, but still was not quite right.

So these days, I'm starting to feel that it is something that is never captured by recordings at all that ultimately defines the difference between live and recorded sound. The Yamaha AvantGrand and the Vivid-CAT-Esoteric-Stealth audio system succeeded in every sonic parameter but one: the intensity of the original sound. Intensity, defined as the sound power per unit area of the radiating surface, is the reason why, even if you could equalize a note played on a flute to have the same spectrum as the note played on a piano at the same sound pressure level, it will still sound different.

Ultimately, therefore, it is perhaps best to just accept that live music and recorded music are two different phenomena. I appropriated the title of this month's "As We See It" from Evan Eisenberg's book-length essay, The Recording Angel: Explorations in Phonography (first edition, McGraw-Hill, 1986; second edition, Yale University Press, 2005), which is essential reading for anyone who, like me, is fascinated by the art of audio recording. Eisenberg's thesis is that any attempt to capture the sound of an original event is doomed to failure, and that stripping a concert from its cultural context by recording only the audio bestows a sterility on the result from which it cannot escape. The recording engineer may be able to pin the butterfly to the disc, but it sure doesn't fly any more. For a recording to make the grade as a work of art, therefore, more is needed than merely darkly echoing the original event.

In Eisenberg's words, "In the great majority of cases, there is no original musical event that a record records or reproduces. Instead, each playing of a given record is an instance of something timeless. The original musical event never occurred; it exists, if it exists anywhere, outside history."