Digital Processor Reviews

Five years after the launch of the audiophile's dream medium, Super Audio CD, the format remains stalled in the market.

Partly this is because of the market's confusion due to the competing DVD-Audio format. Partly this is because of the slow growth of recordings available on SACD, at least in the beginning. Partly this is because some SACDs remastered from older masters aren't of sufficient quality to justify the price premium for an SACD release. Partly this is because the mass market has always valued convenience over sound quality, hence the conquest of the LP by the barely hi-fi cassette and the current success of MP3. And partly this is because the multichannel aspect of the SACD medium (and of DVD-A) doesn't fulfill a need on the part of a sufficiently large number of potential customers: the majority of the relatively small number of customers who appreciate high sound quality are currently satisfied with two-channel, and while the mass market does appreciate multichannel playback, this is currently in a video context only, and those customers are happy with the compromised sound quality offered by Dolby Digital.

But part of the reason is also that, to get the best from SACD, you need a high-performance, hence high-cost player. Inexpensive SACD players tend not to sound any better with SACDs than with CDs, in my admittedly limited experience. But when all the stars align, and an SACD with true hi-rez program is played on an appropriate player, the result is sound quality that exceeds anything that has been available before to audiophiles.

Which brings me to the combination of the Verdi SACD/CD transport and the two-channel Elgar Plus DSD D/A converter, both from English company dCS. The usually taciturn Michael Fremer reviewed this setup in April 2003 (Vol.26 No.4), along with the dCS Purcell D/D processor used to upsample CDs to DSD, and ended up raving about the sound quality: "Instrumental focus, separation, three-dimensionality, and resolution of low-level detail set the Verdi–Elgar Plus apart from other SACD players I've auditioned." He concluded, "Spend some time with the dCS Verdi–Purcell–Elgar Plus. You'll understand how someone with $34,000 to drop might gladly do so for the pleasure of their company." While I did spend a small amount of time with the dCS gear when I performed the measurements to accompany Mikey's review, I gladly welcomed them back into my system in order to review the dCS Verona.

One gentleman from Verona
In professional audio, a whole studio's worth of digital audio gear is slaved to a central word-clock signal, so that A/D and D/A converters, digital audio workstations, and outboard processors all work in sample synchronicity. dCS has long sold a professional master-clock unit for use in such situations, but never thought of offering such a device for use with their consumer products. The domestic component chain is very short compared with a studio's, so it's hard to see where the need for a separate master clock would arise.

Even so, to play back SACDs and CDs with maximum fidelity, dCS requires that three components be used—the aforementioned Verdi, Purcell, and Elgar. Apparently, some Japanese audiophiles experimented with clocking the three units from a central source rather than relying on the transport to provide the system clock, as is usual with separate transports and DACs, or slaving the transport to the DAC, which is theoretically superior. They reported improved sound quality, so dCS looked to see what was going on.

The result of their research was the Verona, priced at $6995 and previewed at the 2004 Consumer Electronics Show. Housed in an enclosure that looks identical to the Purcell upsampler's, the Verona features a highly stable (±1ppm, or parts per million), crystal-controlled oscillator from which is derived a standard TTL word-clock reference signal running at 44.1kHz or 48kHz. This signal is available on five 75 ohm BNC output jacks on the rear panel. There are also three RCA jacks providing a 1V p–p, S/PDIF-formatted datastream with the same clock reference frequency.

There is also an External Reference word-clock input on a BNC jack. This overrides the local crystal oscillators, and will accept either TTL or bipolar clock signals—the latter are available from, for example, a 10MHz GPS receiver or an atomic clock. (The ability to sync to a GPS clock will be useful for occasions where musicians are working on the same recording project despite being geographically separated.) The Verona's input receiver has a pull-in range of ±300ppm, and takes around 12 seconds to lock.

Very unusually, the Verona's clock outputs feature switchable dither. According to a conversation I had last year with dCS's chief engineer, Mike Story, the use of dither (a small random timing offset) avoids the "dead zone" that receivers using a phase-locked loop with a very narrow acceptance window can suffer from. The dither noise—said to be statistically well controlled so that the noise energy averages out to zero in the long term—ensures that the receiving PLL remains locked to the Verona's output with the minimum of jitter. The dither itself is filtered out by the PLL.

Sound: Elgar & Verdi revisited
Before I set up the Verona in my system, I spend some time refamiliarizing myself with the two-channel Verdi–Purcell–Elgar Plus combination, driving the Mark Levinson No.33H monoblocks directly via 10' lengths of Madrigal CZ-Gel balanced interconnect. (The Elgar's remote-controllable volume and selectable maximum output level facilitates this mode of operation.) To play SACDs, the Verdi fed the Elgar via a FireWire link; the Purcell was hooked up to the Verdi's AES/EBU data output and sent the upsampled DSD-formatted data to the Elgar, again via FireWire. The Elgar acted as the clock master, and, usefully, automatically recognized when a CD was being played via the Purcell or an SACD was being played on the Verdi, switching its inputs appropriately.

"Truthful, analytical delivery and [a] lack of tacked-on warmth," wrote Michael Fremer in his April 2003 review of the dCS system. More important, I found, there was a complete absence of the fatiguing high frequencies that are so common with CD playback. This was as true with modern, state-of-the-recording-art, all-DSD recordings such as Tierney Sutton's Something Cool (SACD, Telarc SACD-63548) as it was with classic recordings derived from analog master tapes—Dave Brubeck's Time Out, for example (SACD, Columbia/Legacy 7464-65122-6).

As well as the overall ease to the sound, I could hear into the soundstage to an extraordinary degree. The disposition of the musicians on K622, Antony Michaelson's reading of Mozart's Clarinet Concerto, which I produced last year (SACD, Musical Fidelity MFSACD017, available from this website), sounded vividly realistic, everyone unambiguously in position, but without recorded detail being thrust forward at me.

That isn't to say the dCS stack smoothed over details. Such incidentals as the wayward intonation of Eugene Wright's double bass on "Pick Up Sticks" from Time Out, or Paul Desmond's clams on alto sax on "Take Five" and "Blue Rondo a la Turk" were distressingly obvious to a degree I hadn't experienced before. (It puzzles me why no one involved with the production of this 45-year-old classic was ever bothered enough by these problems to ask for another take.)

Even without the Verona, I quickly became very comfortable with the dCS's sound. I'd almost say it was the best digital source I have had in my home, though the memories of a weekend I spent last June with EMM Labs' SACD transport and DAC6 suggest that it isn't alone at the top of the digital hill. But with respect to the mighty duo of the Mark Levinson No.31.5 and No.30.6, which together have been my reference for CD playback for going on a decade, they were just pushed to one side by SACD played on the dCS combo, even when both went through a conventional preamplifier.