Filed Under: Things that shouldn’t matter
Staring at the Innuos Statement (website), a $13,750 dedicated computer audio server, I’m reminded of a long-ago period in the evolution of digital audio. If you’ll cast your thoughts back to a time when dinosaurs roamed the earth (okay, fine: it was about 10 years ago), the big debate on the Computer Audiophile forums was about jitter. Nothing mattered but jitter. Jitter was the thief of dreams. Jitter was the enemy. Jitter was the thing that killed the sound quality of digital audio systems. If you could minimize jitter – that series of “timing errors” that meant that your digital processing system was not sending/receiving data in as orderly, predictable, and systematic a fashion as was (perhaps) required for optimal handling and playback – then all would be well with the audio universe. Analog would (finally) be dead (again). CDs could be retired (halleluiah). And then – rejoice! rejoice! – we would wander forth from our caves into the full glory, sunshine, and promise of the audio-file audiophile. All praise the Master Control Program!
Turned out, to the shock and horror of just about everyone on Computer Audiophile – and to the surprise of literally no vinyl fan – that it wasn’t quite as simple as all that. Turned out, as with just about every aspect and avenue of high-end audio perfectionistic fetishism, that everything matters. Power. Components. Design. And jitter. (And probably a bazillion other little nonsensical shiznizzles, but really, who’s counting.)
The Innuos ZENith, when I first got it a year or so ago, was a revelation. Plugged directly into a BorderPatrol Analog DAC SE (reviewed here), the sound I heard was remarkably relaxed and fully transparent. That DAC is a bit of a microscope – with almost no processing at all (no upsampling, no oversampling, no filtering), what goes in comes right out. And what I heard was, quite clearly, better than my Mac Mini. In fact, the Zenith handily and unceremoniously unseated every server I’d had a chance to play with, with one sole exception – the Aurender W20, a nearly 5x as expensive dedicated audio server – and in so doing, thereby levered itself firmly into the Part-Time Audiophile Editors’ Choice Awards.
And that was that.
Enter: the Innuos Statement
I wasn’t under any illusions that the ZENith couldn’t be bettered. A cleverly-built server it was, but it was “just” a server. Yes, it uses a custom OS – and so, shouldn’t be as vulnerable to the semi-random requirements a general-purpose OS like Apple’s OS X, iOS, or, heaven forbid, some version of Windows [insert shudder of existential horror, here]. But as with the genre of audiophile cables, I (naïvely) felt that the most glaringly obvious “tweaks” in computer-based audio systems would tend to be entirely negative. Improving the power supply, the clocking, the signal path – all that would surely help, but would those improvements be – again – “glaringly obvious”? Again, this is the server we’re talking about. It’s just a transport. The DAC takes whatever the server gives it and runs it through whatever it runs it through (depending on the architecture of the digital conversion process in question), with the outcome of that being FAR more important than whatever the input was. To put it another way, a good DAC should be able to make diamonds out of dust. Crap signals could (and probably should) be buffered on-board in order to eliminate timing errors (and dropouts, and every other kind of error) that the server or network might introduce. In a modern design, the signal could/would/should then be clocked directly into the digital conversion chips using an extremely precise onboard clock – minimizing, if not
murdering eliminating the Dread Pirate Roberts of Digital Audio, jitter — and then Bob’s your uncle. To put it yet another way, if you’re using a well-designed DAC, the server/transport/“ultimate source” should be irrelevant. The fact that it isn’t, pretty much ever, is annoying. A fully decked-out PS Audio DirectStream, for example, with all of its clocking and buffering and digital awesomeness, is nevertheless very sensitive to the quality of the transport feeding it. And that’s bizarre. But I digress.
After a long series of “teachable moments”, I came to believe that with a good server, good cables, and a revealing DAC, the sound quality would change with any component change. But I also figured that the sound quality would not change much, unless I was changing the DAC. That is, if there would be a non-negative change, I figured that the changes upstream from the DAC would be, at best, minimal.
I’m an idiot. This is known.
Diamond Revision 30th Anniversary USB Cable from Purist Audio Design
I want to take a moment to talk about a USB cable: the Diamond Revision 30th Anniversary USB Cable from Purist Audio Design. It’s USB 2.0 cable, yes. One that comes with all the bells & whistles. Silver wire. Upgraded connectors and cable dielectric. Isolated (as in “physically separated”) data and power legs (but doesn’t require a separate USB connection to pull power). Passive power filtering (ferrite on the power leg, not the data leg). It really is about as over-built and tech-heavy as a USB cable gets, without integrating “active” elements – like a regenerator, a reclocker, or battery-power. In short, this is a nicely made and thoughtfully designed wire, and requires no special degree to use or understand.
… and it sounds great. Feel free to cue up all of your “bits are bits” comments now.
I get it. It shouldn’t matter. Again, in a system with properly designed gear, cables generally aren’t supposed to matter. They’re supposed to be the afterthought. Their impact, if positive, ought to be absolutely minimal; when audible, they are probably negative/subtractive. That’s the Prevailing View. Again. I get it.
In a digital system, a cable is just supposed to transmit “ones” and “zeros”, or more properly, “signal/no signal”. That’s it. Nothing magic. If “enough” signal goes through, the receiver is supposed to get all it needs to reconstruct the signal exactly. What can a cable contribute to that reconstruction? Santa’s good cheer?
Again, I get it. I am also going to take the following as read: if that is your stance, feel free to do a bit more reading. In the meantime, I’ll just point at the fact that digital signaling across a physical medium like a wire is a form of signaling, and is done entirely in the analog domain. The wire matters. The connection matters. The power coming across the wire matters. In short, like everything else in high-end audio – and much to my everlasting annoyance – everything matters.
The problem this stance had to overcome was born of pure experiential data. In short, I tried this cable out on many people. That is, I was pretending to screw around with “something” in the system while swapping only this cable in and out. Over the course of a year, I did this with dozens of people.
Over the course of my “testing”, there were several cables that met with acclaim, but only a very few that folks were able to agree were “the best”. The Diamond Revision 30th Anniversary USB was the best of that group – best improvement to the sound and the easiest to use.
By “improvement”, I mean something like “ease”. That is, there is an easing to the sound when good digital cables are used. Some folks say that this is a lack of edginess, an analog type of smoothing, a removal of something inherent in digital playback that puts your teeth on edge. I don’t necessarily agree – there is nothing lost with the Purist USB. Quite the opposite. Instead, I have that same effect I get with the ZENith server – a sense of relaxation, like the difference between a warmed-up band playing their favorite hits and that same band exploring a brand new tune. Perhaps it’s better to say that it’s as if there’s a fullness to the sound, a wider envelope, more comfort. I tend to call this phenomenon as “organic”, because it puts me in mind of the difference between skin seen with the eye and skin seen under the hand. There’s just something different to being physically and immediately present that photography abstracts you from. I suppose that might mean a timbral fullness, a fuller spectrum of transmission, but I suspect the numbers there would be inconclusive. Instead, I think the closest physical analog I can make is the difference between an amp with a “proper”, fully-spec’d out, well-designed power supply, and one that is “overbuilt”. BorderPatrol amps are like this – you can get one of their power amps with the “proper supply”, and that sound will be heaven wrapped in silk. And then, when you’re ready, you can upgrade that “proper supply”, and begin to enjoy those “wait, what just happened” experiences. And then, when you’re ready, you can upgrade the upgrade, and have those “uncontrollable giggling” episodes that may require pharmaceutical intervention to control. Great cables are very much like this. Shouldn’t matter! The first damn wire was “built to spec” and everything else should be unnecessary, a technical gilding of the lily. But the giggle is right there, nonetheless, mockingly reaching out to fondle your goosebumps as you sit there, paralyzed, saying “See? I told you so.”
The Purist Audio Design Diamond Revision 30th Anniversary USB is now my reference.
One caveat: this cable is over $2k retail – that is, it’s expensive. I’m not going to argue the economics of why this is a good value, but I will instead suggest doing the reasonable thing and borrowing one to try out for yourself. If you have an Innuos server, the experience is recommended.
Find out more at Purist Audio Design: https://www.puristaudiodesign.com/Data/products/usb/usb_30th.html.
Back to the Innuos Statement
So, the long story short is this: I should have been less surprised when I finally got the Innuos Statement from Mark Sossa of Well Pleased A/V. I put it on the rack next to my ZENith and let it warm up for about an hour while we screwed around with iPeng, the interface I’ve been using to control the ZENith, and now, the Statement (more on that, below).
From the outside, The Statement is a very different animal. You have the same swoopy/angular faceplate, but there are now two boxes (usually stacked vertically) instead of one, coupled together by a pair of umbilicals. That second box is a different power supply – a linear PSU with “8 independent rails”:
Partnering once again with Dr. Sean Jacobs, a new power supply architecture was designed exclusively for the Statement. The power supply is distributed between two enclosures, with one enclosure containing the AC/DC conversion stage and the other containing the regulation stage together with the server components.
The advantage is clear: regulated “clean” power traverses a very short path, soldered point to point, with the least resistance and exposure to EMI. This is superior in efficiency compared to a traditional architecture where a separate power supply is used – regulated “clean” power needs to go through more obstacles of internal cabling, output connectors, cable, input connectors and more internal cable before it reaches its destination.
Each critical component of the server receives its own dedicated power supply from one of the 8 independent power rails, further preventing component cross-contamination. Of particular note is the use of a dedicated power supply for each of the Ethernet and USB high-precision OCXO clocks, ensuring these clocks work at their best.
The motherboard is “hardened” against EMI:
The motherboard has been fully EMI-optimised to our own exclusive design by, removing any unnecessary, noise-generating components capable of compromising audio. A fully-customised BIOS ensures the hardware works at its lowest noise without impairing performance. Further EMI treatment is then applied to the motherboard so that EMI is absorbed rather than reflected to other components.
And perhaps the most telling upgrade of all has to do with jitter:
Not fully satisfied with existing USB and Ethernet re-clocking designs, Innuos designed in-house their own boards for this purpose. At the heart of the design are 3 principles: The use of very high-precision OCXO clocks for timing the signals, extreme care in powering important components preventing cross-contamination, and extremely short paths between components.
While the use of OCXO clocks is not new, the way they are powered and the layout of the board has proven to be extremely important for their performance. As an example, dozens of USB controllers were analysed in terms of how they were powered and clocked so that they allowed this control to be done externally. The USB Controller is regulated by no less than 3 ultra-low noise regulators, one per independent voltage. Keeping the clock very close to the controllers, via a dedicated track on the board itself, helps to avoid the losses in signal quality that come with the use of external clock units.
The entire package is rather elegant, which may be a weird thing to say about a computer, but it’s true. The ZENith looks nice. The Statement, with its massive two-tier appearance and custom-cut top-plate, looks expensive. I suppose that’s fair, as the pricing starts at $13,750, and goes up with onboard storage options. I will note that this is a significant cost saving over its primary rival, the Aurender W20SE ($22k), though that saving comes with some limitations — all of the Innuous products are USB-only.
Aside: a random USB rant
It should go without saying, at this point, that USB is the default interface for DACs. There are some notable holdouts, including Berkeley Audio Design (who made one of my very first reference DACs, all those years ago), and PTA favorite Tidal Audio of Germany, which makes ultra-fi loudspeakers (like my reference Piano G2 in a spectacular plum-colored no-fingerprint finish) and matching electronics, including several DACs, none of which have a USB input. I find this annoying.
Yes, I’m sure there are many good reasons for why DACs do not have USB inputs. One of them is, no doubt, that computer audio electronics are tricky enough to manage — having to now answer user questions about drivers and software for computer hardware that is neither predictable or manageable is anathema to a business plan. More, computer audio electronics also tend to evolve faster than amp or preamp tech — which means an increased risk of having “combo” products move quickly out of date; in some cases, this “decay” can move even faster than it takes to release said products in the first place. Vitus Audio and Vinnie Rossi, at various times, have faced this rapid technical obsolescence with their USB interfaces, for example. Of course, these two companies had fully modular solutions, so user upgrades were simple. And given that USB is now fully ubiquitous — and has been for the better part of the last decade — I think it is safe to say that requiring an S/PDIF output on your transport is an unpleasant anachronism. To all DAC makers: please stop doing this.
Listening to the Innuos Statement
Plugging the Innuos Statement into my Bricasti Design M1 Limited was revelatory!
This made me very upset.
The sound quality of the entire system jumped in a way that should not have been possible. Dammit, my brain screamed, the server shouldn’t matter this much! But what I heard was undeniable — musicians were significantly more tactile and now standing on a deeper stage, with separation and nuance far more obvious than with the “regular” ZENith, a server which I had only moments before, been thoroughly happy with. The sound simply “tightened up”, and a diffuseness that I would have sworn moments before did not exist, disappeared. I might have gone a bit slack-jawed.
The ZENith Mk3 retails for $4,249 dollars. The difference between the ZENith and the Innuos Statement is over $9,000. Had I heard the Statement vs the ZENith head-to-head, as I heard it in my own rack, however, I probably would have saved up for the Statement. Hence, my annoyance. The jump in performance quality from one level to another was clear and obvious, and the improvement was across the board. The background was quieter. The dynamics were better. The precision was more spot-on. The experience was more immersive. None of that should have been true, yet, there it was.
Let me be clear: the ZENith Mk 3 server is truly excellent — it beats the literal snot out of the Roon Labs Nucleus (reviewed here). It embarrasses every Mac I’ve ever tried, both with and especially without specialty audio software. The ZENith is a reference-quality solution.
And then there’s the Statement.
Clash of the Titans
My ultra-fi reference for computer transports to-date had been the Aurender W20. That server, with its near-fanatical architecture and tank-like build-quality, set new standards for audio performance in every audio system I introduced it to. It was, in a word, phenomenal.
Head to head with the Innuos Statement, this “old” (5-year-old) Aurender W20 was, unsurprisingly, close. Sonically, both of these boxes “do things” as they improve the overall sound quality, but ultimately it seemed to me as if those were different things. Capturing this involved quite a bit of chasing after ephemera, unfortunately, as the style and quality of the relative effects are damn near impossible to not over-emphasize, and may well be lost to all but the most revealing of audio systems (and I apologize if that sounds elitist, but, oh well). My notes held clues to the following: the Statement pointed toward a more “open” sound while the W20 presented a more “focused” sound. And it’s worth repeating that both of these transports were significantly more engaging, more revealing, and more explosive-sounding than the Mac Mini, the Nucleus, and my previously-satisfying ZENith.
When looking at these two solutions, there is one thing I ought to mention — the control app. The Aurender app, Conductor, is reasonably solid and stable, with a largely intuitive and functional interface. I don’t want to overstate the situation, here, because there’s no such thing as an app-based solution that is really worth jumping up and down about — Conductor is not Roon, for example, which is my head-and-shoulders functionality favorite — but Conductor does allow me to categorize disparate sources (local, network, and streaming), which is good and very helpful. By contrast, Innuos offers no native interface solution. I’m told that Innuos may have an interface in the works, however. Let’s hope that’s so. In the meantime, Innuos users might be tempted to use Roon. The combo is elegant and eminently usable.
I want to mention, however, that the Aurender sounds rather different, depending on the location of the music files used for playback. Oddly, and somewhat perplexingly, using Roon with the Aurender was not as satisfying as directly streaming the various go-to audiophile services (via Qobuz, or Tidal, for example), which, when played directly via the native Conductor app, sounded excellent. Interestingly, streaming those same tunes from a NAS sounded better still, and playing from local storage was easily the best of all, and by a significant margin. Interestingly, this stacked-ranking of performance was paralleled exactly in the Innuos Statement. Playback via Roon was bettered by the streaming services, accessed by the freeware solution iPeng, and using that minimal app to play files stored locally made the sound always richer, fuller, and vastly more immersive.
In the end, I preferred the Innuos Statement. On the one hand, my affinity was clearly for the Innuos soundstage presentation — which, to me, seemed to emphasize and promote openness and clarity, and to a degree that I have never experienced in my home before. On the other hand, it also brought obvious functionality enhancements — the new Innuos Statement just let me do more things, including stream whatever multiple of DSD I cared to, while my old W20 was limited by yesterday’s state-of-the-art. Even with the open question about an interface, the Innuos Statement was the better-performing solution, all around.
If you’re an audio dealer or anyone that demos audio gear for friends and customers, the Statement is what you get so that folks can really hear what everything else in your system is doing and what that system is really capable of. If you’re assembling a system to show off at AXPONA, for example, the Statement is the first thing you should pack.
It plays all file formats. It automatically connected with every (USB-based) DAC I tried it with. And every system I dropped it into, no matter how good, sounded better with the Statement at the front of the chain. Given what I currently know, the Innuos Statement is the best there is.
I will offer a caveat here: I haven’t had the new W20SE in yet. And while that remains the shadow waiting in the wings, it’s worth underlining the fact that that shadow is also $22k, 35% more than the Innuos Statement. At some point, a Thrilla in Manilla is called for. But for now, the Heavy Weight Championship Belt clearly, easily, and unreservedly goes to the Innuos Statement.
Interlude: Innous Phoenix USB Reclocker
In an attempt to un-suck computer-based audio, “reclockers” became all the rage. I remember a bunch of false-starts with some dongles and widgets that actually made things worse, but for me, the Berkeley Audio Alpha USB is when the sound quality via USB took a daring step forward, a trajectory arguably culminating with the dCS Network Bridge. As I mentioned, and speaking about today, I think that converters (S/PDIF to USB) really ought to be unnecessary. HOWEVER (there are always caveats, aren’t there?), there are some good arguments as to why/when a USB to USB reclocker would be an excellent option to pursue. For example — when you already have a server/transport and DAC that you like and you don’t want to shell out — just for a random example — $13,750 to invest in a Statement (and whatever else for a new DAC).
Which is probably why Innuos decided to one-up the category with the Phoenix USB Reclocker.
Here’s what we have on the Innuos Phoenix USB Reclocker ($3,149), announced last Fall:
The Phoenix USB offers in one unit the equivalent of 3 separate components: A USB regenerator, a linear power supply, and a master clock with its own linear power supply.
Innuos applied 3 main design approaches learning from their experience with the Innuos flagship music server, the Statement:
- The USB chip regenerating the signal contains no switching regulators. All 3 independent voltages to the chip originate from an independent linear power supply with further regulation provided by 3 sets of LT3045 regulators.
- The use of a 3ppb OCXO clock running directly at 24MHz and connected via a board track just a couple of inches away from the USB chip. Therefore, no precision is lost within cables and connectors, as is the case when using an external master 10MHz clock with an additional 24MHz clock generator.
- Two independent statement-level linear power supplies, one dedicated to the OCXO clock and the other used for powering the USB chip/5V USB line.
A friend brought me a Phoenix to play with — we promptly put the Phoenix between the ZENith and the Bricasti M1 Limited, wiring things up with excellent USB cables from Purist Audio Design and let ‘er rip.
The results were interesting.
ZENith + Phoenix
In my system, the ZENith + Phoenix combo was completely compelling. In fact, it was so compelling that I think the vast majority of Innuos customers can cheerfully “end” here. My feeling is (again) that unless your system is intensely revealing, the bump from that point to the stand-alone Statement may well be lost. YMMV.
To be more explicit, and personal, I’m kinda on the fence here. On the one hand, the Statement is just simpler and the sound quality is just better. That makes it a powerful tool for a reviewer. On the other hand, the outlay is a significant investment (at least for me) in a device that (as a category) tends to have a shelf life of, say, 5 years (it’s a computer, after all, and 5 years is a long time for a computer to remain relevant much less anywhere near the cutting edge). On the gripping hand, the Phoenix lets me come awful close on sound quality to the pinnacle of the Statement, and does have some flexibility that the Statement does not — mainly, I can use it with whatever computer transport I want, which means mixing and matching for future reviews. At least part of that last thing, however, I don’t think applies to “normal people”. Again, YMMV.
The interesting thing is this: if I had to quantify the improvement the Phoenix brings to the “base” ZENith setup, I’d be hard-pressed to differentiate this pair from the Statement alone, without having them side-by-side. Combined with the ZENith, the Phoenix added all of the same kinds of improvements that the Statement did. There was the same sense of clarity, the same sense of ease, the same sense of depth. Every aspect of the performances I heard was improved incrementally but obviously. My already great system took another clear step up the path to new heights of awesomeness.
ZENith + Phoenix vs Statement
To repeat, I found the ZENith + Phoenix combo to be very close to the Statement level of performance. I suppose that makes sense; the Phoenix design is largely incorporated into the Statement.
I had the clear impression that there was a progression of improvements with these three products. That progression was not linear, however. Draw a line from the ZENith to the Statement, and that line will not intersect with the ZENith + Phoenix — that would require a curve. Put it this way: the overall sound of the system with the Statement in place was clearly on top. That experience was followed rather closely by the system with the ZENith + Phoenix in play. That experience was followed a bit more distantly by the ZENith alone. Put in the other direction: if you have a ZENith, the Phoenix is a big step up; if you have a ZENith and a Phoenix, the Statement is a step up, but a much smaller one than the step from ZENith to ZENith + Phoenix.
Essentially, we hit the “diminishing returns” part of the arc after the Phoenix. You will get better performance with the Statement. But while introducing the Phoenix to the ZENith had me swooning, going from there to the Statement had me in a more reflective sort of mood. I could do better, yes, but I was no longer feeling the desperate urge to throw credit cards at strangers. With the Phoenix in-system, I still know that there is better out there in the world, but I will probably be able to wait till I can get that check from selling my cousin’s kidneys online (which is good, because those transactions are kinda tricky).
And there you have it.
The Innuos Statement is unapologetically great. Unapologetic to my wallet, that is. And while I cannot afford it at the moment, I will probably attempt to do so in some future iteration of my review system. The Statement is an unqualified good thing — it made my system sound better than it ever has, and all things being equal (and me being unaccountably wealthy), there would be an Innuos Statement in my review system full-time, all-the-time. It’s a brilliant tool and one that gets me closer to the music, closer to “the absolute sound” (if you believe in such things) than any computer-based transport ever has. The Innuos Statement is my current reference for State-of-the-Art and handily earns our Editors Choice Award.
For those that are curious, but a bit more budget-conscious — know that the Phoenix USB Reclocker will take you a long, cheerful, and awesome way along that journey to digital audio bliss.
Add a great USB cable (or a pair of them, for use with the Reclocker), like the Purist Audio Design Diamond Revision 30th Anniversary, and Fanny’s your Aunt.
Cue up the choir and thanks be to the Master Control Program! Halleluiah!
For more information, see Innuos online at http://innuos.com.