A Tale of Two Systems Page 3
If it is possible, then, to make valid value judgments concerning a recording-as-art, then, even if there is no absolute reference (you don't know ultimately what the record should sound like), surely it can still be used to make value judgments about hi-fi components? To insist otherwise is to take too narrow a view. The LP grooves and the CD pits have no memory of their genesis: the construction of a reproduced soundstage, with the images of instruments and voices hanging in space, is an illusion. It is as complete an illusion whether it resembles an original acoustic event or whether it springs totally from the genius of the artist and producer.
Consider, for example, a recording that I use as a reference for judging a system's ability to present a tightly focused soundstage. Clannad's "Newgrange," from the album Magical Ring (RCA RCALP 6072), is a typical modern rock recording in that it is totally artificial. Nevertheless, on a good system it possesses considerable depth of field, just as effectively producing the illusion of a consistently well-defined soundstage, for example, as the simon-pure Sheffield Lab recording of The Firebird.
What is important is not the recording's pedigree; rather, it is that the listener hears in the illusion of an image what he or she interprets as a direct sound from an instrument or voice associated with reverberation. When that reverberation is coherently reproduced, then the listener's brain will interpret the soundsource as being set further back in the soundstage than one with less reverberation. A lesser system can destroy the coherence of that reverberation, fouling the correlation between the apparently direct sound and its reflections and therefore reducing or even destroying the sense of depth.
A superlative system can have sufficient resolving power to reveal when the reverberations associated with individual sources within the mix are disparate, giving the listener a multiplicity of views superimposed. When such differences are reduced in magnitude, then, by inspection, the system must be less good. The fact that the recording being listened to never had any prior reality is outside this process. What the critic is listening for is as transparent a window into the illusion of musical reality as possible, and the origins of the recording are, to quite a large extent, irrelevant.
Which brings me at long last to the subject of this essay. Returning my daughter Heather to the UK in September so that she could start her schooling, I took advantage of the trip to visit friends and magazine contributors I hadn't seen for a while. Martin Colloms seemed the same as ever: sentences falling over themselves as he attempted to bring me up to date on where he found himself standing on digital, loudspeaker design, the state of recording, and—most importantly—the sound of his system. His system? Front end was either a much-modified Cambridge CD1 (the original version with three time-shared 14-bit DACs per channel) or a Goldmund Studio turntable, its ST4 arm carrying a Koetsu cartridge. Preamplifier was the British Fidelity MVX, driving two Goldmund Mimesis power amplifiers, while the loudspeakers ostensibly were bi-wired Celestion SL700s.
I say "ostensibly," because these very much were unique speakers. I have asked Martin to go into some detail in a future Stereophile article, but briefly, the drive-unit voice-coils had all been wound with a high-purity solid-core conductor, the tails of which were long enough to act as the speaker cables. The crossovers were placed adjacent to the amplifiers and internally wired with the same cable, as were all the inductors. As far as possible, all the interconnects and internal amplifier and CD-player wiring were also of the same cable, resulting in both a minimum of metal-to-metal contacts and a homogeneity of conductor—"five-nines" pure (99.999%), single-conductor silver, insulated with Teflon.
The sound? Whereas stock Celestions belong in Stereophile's Class B category, Martin's system verged on Class A. Though lightweight, the sound was seamless from the low bass to the highest treble. It was also more musical than I can ever remember hearing in Martin's room. The accuracy of the imaging reminded me of the very first time I heard Quad ESL-63s, but with much better sources and amplification freeing the sound from its chains. Not only were the individual instrumental and vocal images in the soundstage precisely positioned, it was possible to detect different coloration signatures associated with each of those images, also nicely localized. In effect, the accuracy of the imaging allowed you to more easily detect the colorations at the recording end of the chain: you could hear that one voice in an operatic duet had been recorded with a different microphone from the other, for example.
The wealth of detail was almost overpowering: it was like nothing so much as the vista you get looking down on a city from a plane on a clear night. It seemed that whatever recording we played, CD or LP, the end of Martin's listening room opened up into the appropriate acoustic. On went the Nimbus recording of Britten's Frank Bridge Variations (NI 5025): there was the cavernous Great Hall of the University of Birmingham, the site for all those spectacular (if overblown) Louis Fremaux recordings for EMI in the early '70s. (Remember his Massenet Le Cid?) It was skewed a little bit to one side by the UHJ surround-encoding, but it was real! On went the Sheffield Shostakovich Symphony 1 from The Moscow Sessions (CD-26). Now the end of Martin's room was a long, thin hall, the Moscow Philharmonic having an unbelievably deep staging. Appropriately for this issue, on went the Ashkenazy Rachmaninov Symphony 1 on London (411 657-2): there was the Concertgebouw acoustic, atmospheric in its tangibility.