My Quest for a new DAC, Part 2 - Ayre QX-8

As I described in the introductory article a few months ago, I am on a  quest for my next DAC. Over the course of several articles, I'll be describing my evaluation of various candidate DACs that I try out. My requirements and criteria were laid out in the introductory article. 

In Part 1, I evaluated the first candidate, the Mytek Brooklyn DAC+ with Uptone JS-2 power supply. Much as I loved this combination, I looked forward to evaluating several more stellar candidates waiting in the wings. It's a tough job, but somebody has to do it! Did I mention how much I'm enjoying this journey?!

In this article, I will evaluate a DAC whose arrival I have been eagerly awaiting for many months: the Ayre QX-8. As I mentioned in my introduction, I am a huge fan of Ayre DACs. My current DAC, the Ayre Codex, has proved a stellar performer, and I regard the flagship QX-5 Twenty, as a reference DAC, one of the best I've heard. Depending on your point of view, the QX-8 should either extend the qualities of the Codex to new heights, or bring the lofty sonics of the QX-5 to a more affordable price point. Ideally, both. Either way, the QX-8 has big shoes to fill!

Since the QX-8 is a brand new offering, I was thrilled when Ryan Berry, President of Ayre, emailed me to say my evaluation unit was on its way. 

Form and Function

I  found the QX-8 to have similar looks and dimensions as it's bigger brother, the QX-5 Twenty. If anything, the QX-8's front panel is more spartan, with just power and input buttons, a small high-resolution rectangular OLED display, balanced and single-ended headphone jacks, and a selector knob. I'm a sucker for the brushed aluminum finish, and the unit, while undeniably rectangular and boxy, oozes quality in its look and feel.

Ayre has designed the 8-series as a modular platform, so the same chassis can be provisioned to deliver a digital hub (QX-8), a preamp (KX-8), an integrated amplifier (EX-8), and a power amplifier (VX-8). Thus it came as no surprise, when peeking under the QX-8's hood, to find that it has a large volume of empty space, which is necessary to accommodate the other configurations in the 8-Series.

Despite its hefty dimensions, the QX-8 is nowhere as heavy as the QX-5 Twenty, but then it doesn't have the flagship's 4 (count 'em!) separate transformers. As Ryan explained in a CA post:

"A main difference between the QX-5 and QX-8 without going too far into detail are: the QX-5 has our passive line filtering technology (AyreConditioner) built-in while the QX-8 does not.  The QX-5 has 4 separate transformers with different windings used to drive the many parts of the QX-5 (Ethernet, USB, Digital, Analog) and the QX-8 has a single transformer with separate windings to drive these different parts of the circuit.  The DAC chip used to drive the QX-5 is the ES9038PRO while the QX-8 uses the ES9038Q2M.  The QX-5 has custom-made SC-cut crystals from Morion, the QX-8 has more traditional oscillators that we chose as having the best measurements we could find "off the shelf" -- they're still customized for our frequency needs, of course.  PCB materials, components used, and the amount of voltage regulation on the two units also differs based on what we could get into the QX-8's budget as is typical with differently-priced lines...."

In terms of significant specifications, the QX-8 is remarkably similar to the Codex, with some additional improvements and features:

• PCM support up to 32/384, DSD support up to DSD128 over DoP
• fully balanced analog outputs
• linear power supplies
• ESS9038Q2M chip
• Ayre’s Minimum Phase digital filter
• Single-pass 16x oversampling
• Ayre’s Diamond output stage
• Up to six digital inputs: Ethernet, Asynchronous USB, AES/EBU, S/PDIF, Optical (2)
• Bus powered USB input.

Ayre explains its design choices

I asked Ayre to explain what aspects of the QX-8 design contributed to its sound quality:

• The incoming A/C power runs through a filter that we designed to reduce high frequency noise from entering the system.  As with all of our products, the unit is double-insulated with the ground disconnected from the house ground to minimize the effects of other electrical devices in the home negatively impacting the system.

• The transformer used is an EI custom 75VA transformer made by Mercury Magnetics, with separate windings for the analog and digital portions of the circuit.  The power supply is 100% linear with no switching power supplies anywhere in the unit.  We also have the transformer mounted in a way to minimize the mechanical vibrations from being transmitted to the chassis, which helps to improve the sound.

• Power to the circuit is regulated through multiple layers of regulation with critical loads, such as the analog stage, DAC, and clocks handled through discrete, zero-feedback, low-noise regulators.

• Throughout the unit, we use specifically selected materials, like aluminum for the chassis and stainless steel for the hardware to avoid using magnetic materials that we know hurts the sound quality.  Similarly, we also select many of the components used inside the unit and the PCB material that we know work together to create the best sound possible -- learned after hours upon hours of comparative sound tests over the past 25 years.  When we can't find components that meet our criteria, we have parts custom-made for Ayre to our specifications.  This methodology carries from the components used in the circuit to the connectors used to plug in the XLR cables and internal wires and really highlights the attention to detail we put into our designs.
   
• From the inputs, the QX-8 offers a number of configurations that can be ordered to best suit the user’s needs.  We have never liked the idea that a customer has to buy an entire USB section if they have no use for one, but would otherwise have to purchase it just to have an Ethernet input, for example.  This also helps those that are adding to their X-8 series over time avoid having to buy the same input over and over again.  If a person changes their mind later or their configuration changes, they can of course have the module added by us at a later time as well. Finally, as new technologies are released, this modular design allows us to upgrade portions of the circuit without having to replace other parts unnecessarily.   Ayre has always strived to allow users to upgrade their products whenever possible, helping keep their component current now and into the future.

• As with all Ayre products, the QX-8 is a discrete, fully-balanced, zero-feedback design using the diamond buffers similar to those we began using with the Twenty upgrade for the X-5 series, including a diamond buffer dedicated to the balanced headphone outputs.  The fully-balanced and zero-feedback distinctions are important to us as at Ayre, as we know how critical these design elements are to avoid coloring the sound and allowing the unit to becoming fully transparent; this lets the listener enjoy the music as the artist intended it.  When we say our design is fully-balanced, it is more than simply balanced-compatible. From the input to output, every portion of the signal path is truly balanced every step of the way.  Similarly, there is no feedback used anywhere in the circuit.  While feedback certainly can make it easier to achieve respectable measurements, our designs remain no-compromise from the top of the line components to the more affordable lines we offer.  The QX-8 is no exception.

• The word clock output and all inputs, from the USB and Ethernet to the S/PDIF inputs are all galvanically isolated with separate ground planes to prevent noise from external components from getting into the QX-8.  Additionally, the QX-8 features asynchronous USB as well as our patent-pending asynchronous S/PDIF technology that we introduced in the QX-5.  This solves the biggest issue with the S/PDIF interface by eliminating the jitter found in the interface.  This asynchronous feature is bypassable when watching video sources to ensure audio and video sync.

• Separate 44.1kHz and 48kHz based custom TXC oscillators are used so that no conversion from one clock rate to the other is ever done inside the unit.  These clock signals are generated directly adjacent to the ESS9038Q2M DAC, which has most of the chip's standard features bypassed to maximize sound quality by using Ayre's FPGA-implemented minimum-phase digital filters.  The DAC chip supports native DSD conversion, input to the QX-8 via DoP.

• The front display is an OLED display.  As I’m sure you noticed, we placed a timeout feature on the display.  There’s two reasons we do this: first, it extends the life of the display, and second (and more importantly), running an OLED display generates a lot of noise inside any device and is not ideal when listening.  By having the display timeout, we are able to give the user a display capable of showing much more information than our LED displays we’ve used in the past while also ensuring that the display isn’t hurting the sound quality.

• The QX-8 is Qobuz enabled and Roon certified.

Pricing

In addition to the modularity of the 8-series product line, Ayre also provides modularity in configuring the QX-8 "Digital Hub," which is reflected in the price. The QX-8 can be provisioned as:

• digital base (S/PDIF inputs)            $4,450 MSRP (US)
• USB (add Asynchronous USB)        + $500   ($4,950)
• Net (add Ethernet input)               + $500   ($4,950)
• Full (add USB and Ethernet)          + $1000   ($5,450)

Operation

The QX-8, like my Codex, is dead simple to use. And despite my having received an early production unit, it proved rock solid in operation, with no glitches or bugs. There is, of course, a bit more complexity in the connection to the rest of my system, since I wanted to test both the USB and network inputs. As I am a Roon user, and since the QX-8 implements a Roon Ready endpoint, using the QX-8 via its network inputs was a trivial matter of locating it in the Roon Audio settings, enabling it, and then selecting it as the endpoint.

The QX-8 took several hundred hours of burn in to sound its best. I had experienced this previously with my Codex, so I was not surprised. Suffice it to say that the sound quality after about 300 hours of burn in was dramatically better than when it first arrived. Ayre claim up to 500 hours for SQ to stabilize, and I can quite believe it. By the time I started my critical listening, I was well past the 500-hour mark.

Methodology

I also want to clarify my methodology. While this is unambiguously a subjective evaluation, based on my listening impressions, I took a scientific approach to the design of experiments, especially when comparing different DACs. What do I mean by that?

1. Vary one thing only (DAC): When comparing DACs, hold everything else constant. This should be fairly obvious when it comes to the rest of the chain around the DAC. In addition, I used the same cables (power, USB, interconnects, clock, etc) with all the DACs I compared. In this context, I want to give a big shout out to Cardas Audio. I was already a fan of, and using their Clear cables in some parts of my system, but they kindly loaned us more of the same, so I could use the same cables on multiple DACs without an undue amount of cable swapping.
2. Level matching: I had already been using an SPL meter for this, but in the feedback to my last evaluation, I was recommended a method using a 1kHz tone and a digital multi-meter, which I adopted. Interestingly, I found that a perfect level match at 1kHz with the DMM sometimes still required a bit of tweaking for the full frequency spectrum of a song. I won't belabor this, but rest assured I sweated the details to get levels matched, both in my primary headphones setup, and in the speaker-based setup as well.
3. Listening methodology: My comparison methodology was to use a combination of techniques. Sometimes, playing short 10-30-second snippets allowed me to hone in on a particular characteristic or difference. However, I also found it useful to listen to entire tracks, or even entire albums before switching, as that gets to the overall ease, musicality, and fatigue (or lack thereof). Many of the comparisons were conducted as single blind tests, but the bulk were sighted. Ultimately, my opinion of a DAC was based on its sustained use in my system over many weeks. 

Listening Evaluations

One of the reasons I was able to burn in the QX-8 for that long was that I had a string of summer travel that kept me away from my system. My first listening session with the QX-8 after my return also happened to be the first time I'd heard my system in over 3 weeks. It was an intensely satisfying session, and I recall thinking - I've never heard my system sound this good. If you're like me, prolonged absence from your audio system can make your heart grow fonder, so I initially attributed this reaction to, um, unrequited aural love. But that feeling persisted over multiple listening sessions, and I realized it was the QX-8 that was making my system sing.

The QX-8's sound signature is a heady mix of dynamic, non-fatiguing, and smooth, while delivering plenty of resolution and a huge soundstage. I'm not going to go on and on about its sonic characteristics in isolation, as I find it's best to describe these things in a comparative context. Remember - all the DACs I'm evaluating in this series are highly-acclaimed, and they all sound very good indeed. You really have to do comparative listening to discern the differences.

Before I jump into comparisons, let me talk about the configurations I evaluated. Here is a diagram of how the QX-8 integrated into my system:

As I've noted before, and should be clear in the topology diagram, I've invested a great deal of care, effort, and expense to optimize my upstream digital chain. So I'm always curious with every DAC I test whether, perchance, that DAC's inputs are so well isolated that they are immune to these optimizations. In the case of the QX-8, I tested with both USB and Ethernet inputs, and I'll cover each in turn.

USB Input

I conducted a simple experiment, where I compared the sound quality with a direct USB connection (from SE to DAC) vs. my full USB chain, in both cases using Lush USB cable(s).

Verdict: the USB optimizations still matter, and improve sound quality profoundly. This was not a surprise, as it mirrors my experience with previous DACs, but it's always worth a check!

Network (Ethernet) Input

Next I evaluated the sound quality when driving the QX-8 as a Roon endpoint via its Ethernet input. This is about as simple and elegant a configuration as you can imagine, as you just connect the QX-8 to your home network and you're good to go. There were two significant findings here:

1. the presence of the OCXO switch upstream of the QX-8 has a positive impact on sound quality, just as it does in other configurations in my system. I verified this by comparing the SE directly connected to my router, or through the OCXO switch.
2. the optimized USB path to the QX-8 sounded significantly better than the direct network path, even through the OCXO switch. Again, this is not surprising, and is consistent with what I had found some months ago with the QX-5 Twenty, when I had it on loan in my system for a few days. 

I am not trying to confuse people with these findings. Do note that these sound quality (SQ) differences in inputs are due to some extreme upstream optimizations I've done, which are not the norm and are very specific to my setup. The QX-8 sounded excellent even in the more common scenario, where it is directly attached to a server or endpoint, and to a switch or router in the user's network. I am not going to assert that USB is the better input in general, even though it is, in mine. This is something you just need to try in your system and determine for yourself.

Upsampling and Filtering

First, filtering on the QX-8 is a no brainer. There are no filters to select. If you like tweaking filters, may I suggest you take up knitting instead. The good news is that, just like the Codex, the built-in Ayre Minimum Phase filter sounds excellent. No fuss, no muss.

As regards upsampling, I played around with software (SW) upsampling of content to the maximum input rates of PCM384 and DSD128, using Roon. I would have used HQPlayer, but in my current chain, the Zenith SE does not implement an NAA daemon. This is a feature several SE owners have requested, so this may be made available by Innuos in the near future. While there was the occasional track, where I felt the SW upsampled version sounded slightly better, on the whole, I found the QX-8 is just best fed with native sample-rate content.

In short: the QX-8 sounds damn good, no matter how you feed it. Additionally, if you've invested care and effort to optimize your upstream chain, the QX-8 will accept these optimizations gracefully, and scale up to even better sound quality.

Comparison with the Ayre Codex

Until my quest is complete, the Ayre Codex is still my current DAC and my baseline. It only took a short time swapping back and forth between the QX-8 and the Codex to hear the family resemblance. Everything I like about the Codex, the QX-8 does similarly, but better. I've described the Codex as having a lovely midrange, a satisfyingly solid mid-bass, and a smooth, refined treble. Dynamics-wise, it can be a bit polite-sounding, which I attribute to its necessarily small PSU, to fit its small case.

In contrast, the QX-8 sounds far more dynamic and expansive, while retaining the Codex's beguiling tonality. It projects a huge soundstage, and sounds denser and meatier than the Codex. And it sounds even more tonally refined.

I love this album, of which I own several versions, both in PCM and DSD. I listened to the 24/192 PCM version. The first track, Limehouse Blues, starts with  just the ambiance at the Jazzpuben Stampen (Pawnshop)  in Stockholm. The atmosphere and soundscape projected is so much more realistic on the QX-8. Once the song begins, the solidity and density of the ensemble in the soundstage are much more palpable. Brush strokes are more detailed, the timbre of the drums, double basses, and clarinet is better rendered.The vibraphone in the solo, between around 3:00 and 4:00, almost take on a luminous glow, and all sorts of micro-details like the hammer strokes can be heard. Finally, at the end of the solo, the piano chords, accompanied by double bass, thrum with far more energy, speaking to the better dynamics of the QX-8.

Those who know me, know that I'm a fan of Gustav Mahler, and more generally, of large symphonic music. Mahler's symphonies are a tour de force, where large orchestras are marshaled to deliver the range of emotion and dynamics that he demanded for his compositions. That's why the recent release of the Mahler 5th by the quirky Natalia Ensemble piqued my interest. Could a chamber ensemble deliver a compelling performance of this traditionally large orchestral piece? To my ears, they do, and in a most delightful way. Cobra's release of this album is available on Native DSD as both a stereo and a binaural master, which I bought in DSD64.

Listening to the 4th Adagietto movement in the binaural version, the QX-8's ability to render a huge soundstage is evident from the first note. While the binaural soundscape is well rendered on the Codex, it is a bit constricted and flat compared to the QX-8. On the latter, individual instruments are better delineated, and sound more realistic. For example, the pluck of the harp strings has a rounder, more physical sound, commensurate with how it sounds in reality.

For a Codex owner looking to upgrade, the QX-8 has to be on their short-list. It is easy to fall in love with, because it has everything you love about the Codex, and - more, bigger, better! 

Comparison with the Mytek Brooklyn DAC+ and Uptone JS-2 PSU

The Mytek Brooklyn DAC+ powered by the Uptone JS-2 was the subject of my last evaluation, and had established itself as the frontrunner in my DAC quest, having edged past the Codex. How did it fare compared to the QX-8?

The thumping drum strokes - death knells - at the beginning of the last movement of Mahler's 10 (Thomas Dausgaard, Seattle Symphony, 24/96) have become a good test for dynamics in my system. With the help of the JS-2, the Brooklyn DAC+ rendered these very convincingly. The QX-8 rendered these with equal, if not better, dynamics. But continuing into the movement, it surged ahead in several areas. The QX-8 had a bigger and deeper soundstage. This seems to be one of its enduring strengths. Beyond that, individual instruments were easier to distinguish, and had better dimension to them. The timbre and micro detail from the flute and oboe, to name a couple, was better on the QX-8. Tonally, the QX-8 sounded calmer and more relaxed, without in any way giving up accuracy or resolution.

I'm a sucker for Dixieland jazz! And I make it a point to visit the Preservation Hall whenever I'm in New Orleans. I bought this CD on a visit many years ago, but I hadn't listened to it in quite a while. For this comparison, I fired up Ti-pi Ti-pi Tin, one of many great songs on this album. It sounded wonderful on both DACs. But over time, differences emerged. Relative to the Brooklyn, the QX-8 sounded a tad less strident, a tad less fatiguing when the brass kicked in. With the QX-8, the singer's voice emerged from a blacker background and the clarinet had a smoother, rounder, and to these ears, more pleasing tone. It rendered the atmosphere and ambiance of the hall more palpably than the Brooklyn, which I attribute back to its excellent imaging.

None of what I have written detracts from how good the Brooklyn DAC+/JS-2 combo is. The QX-8 is just even better. This is what I love about this journey. Just when you think it can't get any better - it does!

Comparison with Mytek Brooklyn DAC+ and Paul Hynes SR-7 DRXL PSU

Paul Hynes is an artisanal PSU (power supply unit) builder, based in the wilds of the Outer Hebrides. I like to picture him in his lair, alone on a wind-swept moor, slaving over his timeless creations, while the wind howls, and the sleet pelts. The SR-7 is his flagship. Sadly, he is winding down - yes, bad pun intended - his custom PSU enterprise, so finding these PSUs will be increasingly difficult for new buyers.

A few weeks after the publication of my Brooklyn DAC+ evaluation, my friend Eric took delivery of his Paul Hynes SR-7 power supply. In the world of standalone PSUs, this is akin to a Ferrari. Everywhere we tried this PSU, even when supplanting existing PSUs of very high quality, the improvement was astonishing. Naturally, I grabbed the opportunity to borrow it, evaluate it powering the Brooklyn DAC+, and compare it with the QX-8.

Now this is when things got very interesting! This match up was very, very close. The Brooklyn DAC+/SR-7 combo almost, but not quite, closed the gap on timbre and refined tonality.  The combo actually pulled ahead on dynamics, which is not surprising. The QX-8 has a single 75VA transformer driving all the internal power supplies. This particular SR-7 has a 250VA transformer, driving 2 rails. While each rail was rated at 6 amps, it has a massive amount of headroom. I suspect strongly that this is what drove the superior dynamics. 

In the end, it just came down to what pleased my ears more, and I felt the QX-8's expansive imaging carried the day.

So if this comparison is moot, due to the vanishing availability of the SR-7, why bring it up? Because it highlights the role of power supplies in a DAC's sonic performance. Many people obsess about, and develop strong opinions about a DAC's D/A converter design philosophy, be it R2R, SDM, chip-based, chipless using FPGAs, ring DACs, pulse DACs, you name it. I don't obsess about these, and am quite agnostic. There are stellar DACs in all these categories. A DAC's performance is a union of many things: the input interfaces, the internal clocks, the D/A conversion itself, filter design, the analog stage design and the power supply, to name a few. Everything matters, but certainly the power supply quality matters a lot. 

Wait - what about MQA?

As you may recall from my introductory article, my listed requirements for my next DAC included full MQA decoding as a desired feature. And as I wrote in the Brooklyn DAC+ evaluation, I was not blown away by MQA on Tidal. Yes, some tracks did sound excellent, perhaps even slightly better than their high-res counterparts. But I was unimpressed with the selection of good classical music on Tidal Masters. The kicker for me was the introduction of MQA Core decoding (the so-called "first unfold") in Roon.

I compared MQA tracks played on the QX-8 with just the Core decode, with the Brooklyn DAC+ with the full MQA decode. I won't belabor this track by track, but in general I felt the HW superiority of the QX-8 over the Brooklyn DAC+ carried the day. In other words, the benefit of full MQA decoding on the Brooklyn vs. just core decoding on the QX-8 was not enough to overcome the inherent HW differences I was hearing between both DACs. Even on the Brooklyn, the incremental benefit of full MQA decoding vs. core decoding was subtle at best. On many tracks, I was hard pressed to hear the improvement.

Given these impressions, and the impending arrival of Qobuz streaming, possibly with Roon integration, of high-resolution content in the US, I am dropping full MQA decoding as a requirement for my next DAC. To the extent that I might continue to listen to MQA content on Tidal, core decoding on Roon is satisfying enough for me.

Comparison with the Schiit Yggdrasil

For this comparison, the QX-8 and I took an excursion to my friend Eric's speaker-based setup. This was also a good reality check to complement my headphone-based setup, as it allowed me to validate my findings in a very different speaker-based environment. Eric's system comprises an Innuos Zenith SE server, driving a chain of SOtM tX-USBultra and dX-USB HD Ultra to the DAC via AES/EBU. He too uses a Mutec Ref-10 reference clock on the SOtM boxes. In our comparison, we drove both DACs via AES, which is another difference from my home setup, which is all USB. The analog chain comprises an Audio Research Reference 6 preamp, Hegel H30 power amp, and Magnepan 3.7i speakers, with dual Rhytmik F12g subwoofers. His Yggy unit does not have the new Gen 5 USB stage, nor the new Analog 2 upgrade. Last he checked, he was #98 on the upgrade queue!

Another thing I like about this field trip, and Eric's setup, is that it allows us to do SPL level matching much more easily. And finally, in these comparisons, we mostly use single blind tests on each other.

On Nightingale, from Norah Jones' debut album (Come Away with Me, 16/44.1), the percussive guitar work at the beginning of the piece had better texture and resolution on the QX-8. It also delivered a more coherent soundstage, with better instrument localization. Tonally, the QX-8 sounded smoother, more refined, and a tad darker, with Jones's voice sounding more natural.

In the Rondo-Finale of Mahler's 5th (Minnesota Orchestra, Osmo Vänskä, 24/96), the QX-8 pulled ahead from the first French horn, projecting a wider and deeper image. Here again, the instruments were better resolved, with more texture and a smoother, refined tone. Perhaps surprisingly, given the Yggy is a beast at dynamics, the QX-8 was able to match it.

Overall, this was the first time I'd heard the Yggy comprehensively outclassed in this system, which is a testament both to how good the Yggy is, and the quality of the QX-8.

Comparison with the Ayre QX-5 Twenty

Just as I was putting the final touches on this article, my local Ayre dealer showed up with his demo QX-5 Twenty in hand. He was curious to hear the QX-8, and so we got a chance to compare these two siblings.

Aural memory is unreliable, and I certainly don't trust mine. I can only do meaningful comparisons when I have both units in hand, and can switch back and forth. I had been so impressed with the QX-8 that I had allowed the thought to creep into my head that it may well be close to the level of the QX-5. Uh-uh. Good as the QX-8 is, the QX-5 takes it to yet another level. The sibling order is maintained! Here again, the family resemblance is strong.

Where the QX-5 pulls ahead are in dynamics, bass, resolution, and refinement. Using a single example, Ti-pi Ti-pi Tin on the QX-5 was a revelation. The drums and basses were rock solid, the clarinet and trumpet playing together were so much easier to disambiguate. And suddenly, I was noticing all the individual instruments that were harmonizing with the two primaries as never before.

So yes, the big brother showed the middle brother who's boss, but so what? The QX-8, at a little over half the price, delivers most of that magic in its own right.

Summary

Being a middle child is tough. It's even tougher when your older sibling is an Olympic champion, while the younger is a precocious prodigy. Carving out a niche in the middle of a product line is a challenge. With the QX-8, Ayre had a daunting task, and a tricky needle to thread - to create an offering that was substantially better than the Codex, while not eclipsing the QX-5 Twenty. I would say they've just about nailed it! One can quibble all day about the tradeoffs and design choices, but in the end, the result has to be evaluated as delivered. And boy, have they delivered.

The QX-8's key strengths are an ability to portray an expansive soundstage, combined with a refined tonality, and ample dynamics - more than I expected, given its modest single 75VA transformer. Compared to the Codex, these represent a substantial improvement. When traversing the DAC price/performance curve from ~$2k to ~$5k, we are past the knee of the curve, and well into diminishing returns territory. So it is a pleasant surprise to see how much more performance that extra outlay buys you. Not only do you get an extra-large dollop of increased SQ, it comes with enhanced functionality in the form of additional digital and network inputs, not to mention Roon endpoint support.

 As I said in my introductory article: 
"modern DACs almost universally sound good. ... the question will not be whether they sound good or bad, but whether they speak to my soul. My focus will be on synergy - with my system, my ears, and my preferences."

The QX-8 certainly spoke to my soul.  It is now the DAC to beat in my "quest!" 

If you're in the market for a DAC in the $5k price range, I highly recommend you give the QX-8 a careful audition.

Evaluation Unit Details

Here are the details of the evaluation unit:

• Serial # 005
• Sys FW 00.00.07
• DSP FW 01.00.03
• Net FW 3.05.15

Associated Equipment

Music Server:              Innuos Zenith Mk II SE
Headphone Amplifier:   Cavalli Liquid Gold
Headphones:               Sennheiser HD800 (Super DuPont Mod), Audioquest Nighhawk
USB Regenerator:        SOtM tX-USBultra
Ethernet Switch:          The Linear Solution OCXO switch 
Reference Clock:          Mutec Ref 10 10MHz clock driving the tX-USBultra
Power supplies:            Utpone LPS-1.2 for switch, Paul Hynes SR-4 for tX-USBultra
Power Details:              Dedicated 30A 6AWG AC circuit, PS Audio P5 PerfectWave Regenerator

Power Cables:               PS Audio AC-12 (wall to P5), Pangea AC-9SE MkII (Cavalli Amp), Cardas Golden Cross (Zenith SE),  Pangea AC-14XL (Mutec Ref-10), Pangea AC-14SE MkII to all PSUs, Cardas Clear Power to all DACs under test
USB cables:                  Phasure Lush USB
AES/EBU cables:           Cardas Clear
Clock cables:                 Habst 5N Cryo Pure Silver
Ethernet cables:            SOtM dCBL-Cat7 (switch to SE), TLS cable (switch to QX-8)
DC cables:                    Audio Sensibility Signature Silver (LPS-1.2), Paul Hynes DC3FSXLR fine silver (SR-4)
Interconnects:              Cardas Clear XLR balanced (DAC to Amp)
Headphone cables:        Cardas Clear balanced and SE cables for HD800

Acknowledgments

Many thanks to Cardas Audio for providing a full loom of Cardas Clear cables to allow identical cables to be used for all the DACs under test!