This review page is supported in part by the sponsor whose ad is displayed above

Reviewer: Srajan Ebaen
Source: Zanden Audio Model 2000P/5000S; Opera Audio Reference 2.2 Linear
Preamp/Integrated: ModWright SWL 9.0SE; Music First Audio Passive Magnetic; Hyperion Sound BEC-P25T

EQ: Rane PEQ55 active merely below 40Hz
Amp: 2 x Audiosector Patek SE; Yamamoto A-08S; Canary Audio CA-308s; FirstWatt F3 & F1
Headphones: AKG K-1000 w. hard-wired Stefan AudioArt harness; audio-technica W-1000
Speakers: Zu Cable Definition Pro; Anthony Gallo Acoustics Ref 3.1

Cables: Zanden Audio proprietary I²S cable, Zu Cable Varial, Gede, Libtech and Ibis; Stealth Audio Cable Indra, MetaCarbon & NanoFiber [on loan]; SilverFi interconnects; Crystal Cable Reference power cords; double cryo'd Acrolink with Furutech UK plug between wall and transformer
Stands: 2 x Grand Prix Audio Monaco Modular four-tier
Powerline conditioning: 2 x Walker Audio Velocitor S; custom AudioSector 1.5KV Plitron step-down transformer with balanced power output option
Sundry accessories: GPA Formula Carbon/Kevlar shelf for transport; GPA Apex footers underneath stand, DAC and amp; Walker Audio Extreme SST on all connections; Walker Audio Vivid CD cleaner; Furutech RD-2 CD demagnetizer
Room size: 16' w x 21' d x 9' h in short-wall setup, with openly adjoining 15' x 35' living room

Review Component Retail: $3,000

Reclocking. The new upsampling?
"We don't really understand why it works. We merely know that it does." [Something to that exact effect was said by dCS, the famed British pro firm whose engineers first introduced upsampling. Very much to their credit, they confessed at the time that they couldn't yet mathematically explain why the obvious sonic improvements.]

"Rubidium is bad for the sound." [That was uttered by Yamada-San of Zanden Audio in a personal e-mail to me and by way of differentiating his master clock oscillator from Esoteric's publicized approach.]

I do understand the basic principle of the atomic master clock in Colorado. It sets the time for the US with nearly unthinkable precision. Within any human life span, that precision is certainly well beyond meaningful deviation. However, I do not really understand Yamada-San's disdain for Rubidium. Nor do I understand why upsampling works. Sure, I'm familiar with the argument of pushing aliasing distortion into ultrasonic bands. That allows for the use of shallower, phase-friendlier reconstruction filters. But does that really explain things? Let's be honest and illustrate our fundamental lack of grasp on these matters. Consider this ingredient list of Great Shakes' Herb Mix, recently spotted at Papantoniou's in Chlorakas, Cyprus: "Coarse rock salt, spice, flavourants, vegetable oil and antioxidant."

Where the heck do "flavourants" belong in the context of explaining what's in a spice mix? In the shitter, that's where. It's convenient to wield popular key phrases for explanations that sound impressive at first. But look more closely. Those big fancy mouthful words don't explain diddly. Rather, they require that you first explain and define them. For that matter and in our example, "spice" is no more telling than "flavourants". Both are empty meaningless terms. They convey exactly all of nothing.

For today's purposes, it's probably best to think of reclocking as the utter absence of even the most subtle flavourants. As card-carrying audiophiles, our task is to discern the very subtle spice mixture of music and its perpetual play - how those ingredients intermingle down to the nano dimension. Vital to that pursuit then becomes that the container in which this mixture is served up be free of its own taste. In digital, the claimed higher precision of reclocking (in kitchen speak, its "disinfected stainless steel" container) should work in that most subtle realm of musically relevant data. Ambient cues. Dimensional context. Harmonic finesse. The finest hues and whispers. Overtones should line up properly to remain integrated with the fundamentals. Timbres should be rendered more faithfully, tones more fully fleshed out. A least so theory predicts. But predicting and affixing cause/effect relationships can become a predicament.

You see, non-upsampling, non-filtering converters from the likes of Audio Note for example work beautifully as well. In fact and to my ears -- and when implemented at Yamada-San's Zanden level as one I'm intimately familiar with -- their lack of whiz-bang math arguably works better than a lot of those terribly impressive upsampling specs. Those specs surely appeal to homo testiculus' fascination with speed. Higher numbers seem to imply performance advantages. Digital turbo boost as it were. However, the X-03's strong showing against the Zanden becomes a reverse reminder. Digital peak performance is possible using conventional number crunching just as it is possible using none of it. With this simultaneous validity of opposing methodologies, nothing profound really gets explained as to why and how these digital audio matters operate.

The G25U as an external upsampler/reclocker crunches beyond what the 192kHz X-03 already does on its own. Well, in a way. For $3,000, Mr. G duplicates what Zanden's external clock link between the Model 2000p transport and Model 5000s converter accomplishes when its I-squared-S Ethernet connection is invoked. The rationale for reclocking is the banishment of jitter. We want jitter so far right of the decimal point that it completely eludes detection by that amazingly keen bio computer we call our human ear/brain apparatus. Or so we're told. Jitter is a catch phrase for digital timing errors. dCS, Zanden and Esoteric all champion external reclockers as the ultimate solution for eliminating those timing errors. Esoteric's top implementation in the G-0s relies on the aforementioned "atomic" Rubidium clock. Like a robotic metronome, it sets the rock-steady pulse against which the binary data stream is synchronized and locked at 0.05ppB for up to three devices (and yes, that's part per billion measured over one month's precision variance). Think $13,000 for that bit of hi-tech magic. Ouch.

(By the way, discovered in 1861, Rubidium "can be liquid at ambient temperature but only on a hot day given that its melting point is about 40°C. It is a soft, silvery-white metallic element of the alkali metals group. It is one of the most electropositive and alkaline elements. It ignites spontaneously in air and reacts violently with water, setting fire to the liberated hydrogen. As with all the other alkali metals, it forms amalgams with mercury. It alloys with gold, caesium, sodium and potassium. It colors a flame yellowish-violet.")

By comparison to the $13K G-0s above, the G25U's $3K asking price is chopped liver. Chicken feed even. True, in the grander scheme of things, it isn't. Either. Far from it. But in the context of what comparable reclockers go for, Mister G is genuinely germane to lesser wallets. The X-03 of course already performs its own fixed 192kHz upsampling (that it could be altered via the stop button protocol is a myth). That particular function of the 25 now becomes redundant. The real ticket in this context is the high-precision crystal oscillator. It's said to be accurate to 1 part per million. More precision, less error, performance closer to the mic feed? That's the inherent promise. It's got nothing to do with more data as the 24-bit/192kHz-and-higher upsampling brigade often erroneously believes. There's no retrievable data beyond 16-bit/44.1kHz. Data density is forever fixed by the RedBook format. It's got to do with less offset or skewing so that the same old data lines up more correctly. 1pmm more correctly to be precise. 12 o'clock noon now happens at precisely 12 o'clock noon, not a fraction of a millisecond earlier or later. Well, the audio equivalent thereof. That certainly does ring bloody impressive on paper. Would it be audible though? In fact, is our hearing even keen enough to worry about 1 part in a million?

That's the wrong question to ask. To be relevant, we'd have to know the jitter spec of the X-03 solo. Turns out, that's "3ppm whereas the UX-3 which adds video is rated by itself at +/- 10 ppm" according to Mark Gurvey. A performance magnitude of x 10 as exists between the universal machine au nature and reclocked seems feasibly audible. Perhaps. But two parts per million between the X-03 and its turbo-charged iteration? Then again, it's a bit hard to imagine that a company of Esoteric's cachet would introduce a $3,000 machine that didn't offer demonstrable effects when leashed as intended to its $7,500 standalone CDP. Nor that it would make exactly that combination available for review if it weren't. These thoughts somersaulted through my mind while I considered my disinclination to burden the audiophile worry lexicon with yet another ultimately empty flavourant or spice whopper. You see, reclocking could so easily become the new upsampling. The misinformation and hype that's been heaped on upsampling over the years is testament to how audiophiles are addicted to and led astray by baloney and catchy phrases. Especially numbers can assume a strange life of their own and suggest things that often don't matter. 1pmm. 0.5ppB. Does yours measure up? Will your system suddenly be passé and declassé if you aren't reclocking your data? Will you be excommunicated from the chic in crowd of HiFi?

I simply don't know. If the G25U leashed to the X-03 makes the latter sound better, Esoteric is on to real science and our ears are unbelievably keen to resolve on such a microscopic level. Whether Esoteric's is the only feasible approach to push a digital piece to the next performance level (said piece would have to be endowed with a clock-link input to accept the external reclocker's output); whether their approach is of particular relevance primarily to Esoteric's X-Series; whether other approaches elsewhere might accomplish something similar or equivalent... those are all fair questions for which I presently won't have the answers. One thing is clear. Esoteric really believes in reclocking. Their new AZ-1 [above] in the new Elegance Series is a 45wpc Class D integrated amplifier and reclocker in a single chassis. Conceptually, I find the logic very persuasive that a reduction of timing errors should equate to better alignment of the inner fabric of music. Consequently, it should stabilize and hence reveal the very smallest of cues which otherwise get obliterated, smeared and smudged out as stains. I'm simply in awe at the prospect that humans should be able to hear down to 2 parts per million (or even billion). "It will be subtle but you'll hear it once you take it out again". That's Mark Gurvey's response when I brought up that - um, precise point about today's assignment.

In the end of course, speculation, conceptualization or the fancy graphs and plots the measurement crowd loves to invoke have no relevance if listening doesn't prove things out. Thankfully, all of us have been born with the requisite test tools. The only thing sad about the whole matter is how many self-proclaimed 'philes don't seem to trust their own ears. They prefer instead to rely on other ears or their own eyes (specs and charts). They act as though listening to music with their ears shut and eyes wide open was even possible. While a conjugal visit with the G-0s is proposed for the future if and when Mark Gurvey can scare up a travelling show loaner to perhaps demonstrate the audible difference between parts per million and billion, today's assignment will concern itself over a mere two parts. Per million. After all, that's all the jitter spec that separates the X-03 from the reclocker. Consider me both skeptical and curious. In the extreme...