My Quest for a New DAC, Part 4 - iFi Pro iDSD (and friends)

• Introduction to the Series
• Part 1: Mytek Brooklyn DAC+ with Uptone JS-2 Power Supply
• Part 2: Ayre QX-8
• Part 3: Denafrips Terminator

Next up in my series is the very intriguing iFi Pro iDSD DAC (US MSRP $2499). I’ve had my eye on this puppy for the almost 2 years it’s been teased about and hinted at by iFi, and when concrete details finally started to emerge this spring, I knew this had to be on my audition list.

On the face of it, iFi acts and sounds like a consumer, lifestyle product line, complete with breathless and somewhat hyperbolic marketing, featuring bland phrases with ©, ®, or ™ after them. But marketing does not detract from engineering, and Computer Audiophiles have long admired and lauded iFi’s notable products like the micro DSD Black Label, for its outstanding sound quality at its price point.

Features

The Pro iDSD, while still branded iFi, boasts a feature set and quality that would qualify it as a high-end DAC. Rather than describe features that are table stakes for modern DACs, I’ll just point out the things that caught my eye that really made me want to get the Pro iDSD in-house to evaluate. These are:

• Input sample rate support up to 768kHz PCM and DSD512 native (DSD256 DoP) via USB
• Galvanic isolation on all inputs
• Built In DSD512 and DSD1024 upsampling with a Crysopeia FPGA Digital Engine
• A 10MHz reference clock input 
• Support for external PSU - 9v to 18v DC power supply
• Selectable solid-state and tube analog stages
• Selectable bitperfect and upsampling filters
• MQA decoding (will be delivered in future firmware update)

As before, I’ll be focusing the bulk of my evaluation with the USB input, with a detour to AES for comparisons with the Schiit Yggdrasil.

For completeness, I’ll mention a couple of features I did not evaluate: the Pro iDSD supports Ethernet and WiFi input, using DLNA/UPnP and Airplay, but no Roon endpoint support. Since my setup is Roon-based, and since I’ve optimized the USB chain, I didn’t evaluate this path.

The Pro iDSD also has a built-in headphone amp, but delivers balanced output through a TRRS port, for which I did not have an adapter. In any case, since my system already has an endgame headphone amp in the Cavalli Liquid Gold, my main interest in the Pro iDSD was as a DAC.

Overarching theme: Scalability

The uniquely intriguing thing about this iFi flagship is that - unlike DACs like the Denafrips Terminator, or the Ayre QX-8, where all or most of the choices have been made by the manufacturer - the Pro iDSD has 2 important degrees of freedom along which it can potentially scale: the power supply and the clock. And these two components are hugely important in a DAC. By scaling, I refer to the ability to enhance sound quality.

Don’t get me wrong - there are plenty of other ways one can scale a DAC: with the choice of cables, by optimizing the upstream digital chain, by optimizing the AC power, and by upstream software upsampling, to name just a few. In addition, the Ayre can be purchased with or without USB and network inputs. But having the freedom to vary PSUs and supply a low phase-noise reference clock begged the question: how far up can this scale? I’ll get into this in more detail as we dig into each aspect.

Note that I said potentially scale. Take the power supply. iFi’s website says this about the power supply design:

“All incoming DC is converted to a high-frequency waveform then rectified and filtered by a choke input capacitor filter. This produces a first-level DC bus from which all further voltages are derived. The circuit also generates a galvanically-isolated power supply voltage for the USB input circuitry.

The digital section is powered by a bank of Super Capacitors totaling 6.6 Farad (6,600,000uF). iFi uses Elna Dynacap DZ (TM) Super capacitors because they have a 400 times lower internal impedance than common grades of super capacitors. Individual low-noise TI LDO Regulators with local LC filtering provide the final low-noise power for all individual digital sections, a total of six individual regulators cover Clock, SPDIF Input and the DAC’s digital section. For the analogue stage (especially the tubes) higher voltages are needed.

The whole stage effectively operates on a 60V rail offering massive potential dynamic range.”

That’s a lot of processing between the DC input to the actual circuitry. Ultimately, the actual “power supply” of a unit is the combination of the external PSU and the internal power stages (regulators, super capacitors included). So the question I wanted to explore was - does supplying external power from two PSUs of increasing quality to which I have access, the Uptone JS-2 and Paul Hynes SR-7, cause sound quality to improve (i.e. scale)?

As regards clocks, here is iFi’s description:

“Internal space limitation of Pro iDSD allowed us to equip it with the best clock and tech we could fit inside, but there’s room to maneuver still. If one has a really good (we mean REALLY good – Stanford Research Systems PERF10 or better) external clock, iFi audio has provided the option to try it out with Pro iDSD and go with what sounds best.”
 

Since I own a Mutec Ref-10 reference clock, whose phase noise specifications handily exceed those of the aforementioned Stanford Perf10, does connecting a Ref-10 to the Pro iDSD cause sound quality to improve (i.e. scale)?

Of course, inquiring minds want to know what exactly the external reference clock impacts. Does this only affect the sample rate clocks in the DAC? Or does it extend to the “system” clocks, such as USB, Ethernet, and others. iFi only pointed me to their Global Master Timing® clock paper on the AMR website, which didn’t really address this issue.

Ultimately, of course, the only thing that matters is - how does it sound?

Other Questions

Scalable PSU and clock are certainly the most exciting items on the feature set, but there are other unique attributes of the Pro iDSD that I wanted to explore. 

1. There is a choice of analog output stages: Solid-state, Tube, and Tube+, each with its own sound signature.
2. The Pro iDSD is rather unique in that is provides an internal implementation in FPGA of a DSD upsampler, with two settings - DSD512, and DSD1024.

Here is how iFi responded to my questions about their filters and DSD Remastering options.

a) You provide 5 different filter options: Bit-Perfect (BP), Bit-Perfect+ (BP+), Gibbs Transient Optimized (GTO), Apodizing (APO), and Transient Aligned (TA). These only apply to PCM inputs, correct? 

BP & BP+ apply only to PCM Inputs, they are not available for DSD input, as "Bitperfect" rendering of DSD implies to leave the original DSD Sample Rate.

The standard operation is to oversample PCM to 16X PCM (705/768kHz) and to leave DSD untouched, however BP/BP+ pass PCM untouched (non-oversampling/zero-oversampling etc.) without applying digital filtering.

GTO, APO & TA apply equally to DSD or PCM Inputs if DSD upsampling is selected.

b) What about DSD64, DSD128, DSD256, and DSD512 inputs? What filter options exist?

If the DSD Remaster option is not enabled, all DSD processing is bitperfect. If DSD512/1024 Remaster is enabled GTO, APO & TA Filters are available for DSD64-DSD256. DSD512 always passes unprocessed (as it is already upsampled as no recording hardware exists).

c) Could you clarify how the DSD Remastering - to DSD512 or DSD1204 - interacts with the Digital filter selection?

DSD signals if DSD Remaster is selected, are first converted single step into 705/768kHz/32 Bit PCM (which has more than enough time domain and amplitude domain resolution to represent even DSD256) using the selected filter to remove signal components above 352/384kHz  and then converted to DSD512/1024. 

In order to apply any manipulation to DSD (be it level changes in the digital domain like fade in/out or more complex processing) DSD must be converted to PCM. In most DAC Chip's this conversion is build in in hardware.

Please review this little set of notes on DSD/PCM and various related issues: LINK

d) It would be helpful to describe the flow in the context of a concrete example. Say the source stream is 16/44.1 PCM, selected filter is GTO, and DSD Remastering is set to DSD1024. What is the processing flow from input to output? 

PCM Output: 16/44.1 -> 705.6kHz with GTO Digital Filter
DSD1024 Output: 16/44.1 -> 705.6kHz with GTO Digital Filter -> DSD1024

Listening Experiments 

Due to the plethora of scalability options and features, my listening and comparisons were interleaved, so here’s a roadmap to help navigate the rest of this article. The Pro iDSD was evaluated in 4 distinct configurations, each of which essentially represents a different DAC combo.

The 4 configurations evaluated were:

1. iFi Pro iDSD out of the box, powered by the included iFi iPower SMPS, a.k.a. Pro iDSD (stock),
2. iFi Pro iDSD with Uptone JS-2 PSU, a.k.a. the Pro iDSD/JS-2 combo,
3. iFi Pro iDSD with Mutec Ref-10 reference clock, and Uptone JS-2 PSU, a.k.a. the Pro iDSD/Ref-10/JS-2 combo
4. iFi Pro iDSD with Mutec Ref-10 reference clock, and Paul Hynes SR-7 PSU, a.k.a. the Pro iDSD/Ref-10/SR-7 combo.

Initial Listening

My Pro iDSD unit arrived very shortly after a couple of other review units, so it had many weeks of running in before I got a chance to listen critically. Even with its stock PSU, the sound quality was impressive.  What I really liked - without any comparisons, just listening - were the smoothness and neutrality. There was plenty of detail, dynamics, and air. This was very promising, given that we hadn’t even started on our upward scaling journey.

But before doing any scaling, it was time to compare what I was hearing on the Pro iDSD with the other DACs in my possession in this price range - my current baseline, the Ayre Codex (MSRP $1995) and the Mytek Brooklyn DAC+ (MSRP $2195).

Comparison: Pro iDSD (stock) with Ayre Codex and Brooklyn DAC+ (stock)

I won’t describe this comparison in detail, other than to say these 3 DACs are each outstanding in what they do. Both the Pro iDSD and the Brooklyn DAC+ have the edge on punchiness and dynamics over the Codex, but the latter has a beguiling tonal sweetness that hooked me when I first bought it, and that I still love. The Pro iDSD and the Brooklyn DAC+ are very similar. While the Pro iDSD does not (yet) do MQA, it does have the more impressive feature set. Sound signature wise, I found the Pro iDSD a bit smoother sounding, while the Brooklyn extracted a smidge more detail. I’ll come back to this in subsequent sections. 

Bottom line: out of the box, all 3 of these DACs made a compelling case for themselves, and choosing between them comes down to listening preferences and features.

Improvement with JS-2 power supply

As I had done during my evaluation of the Brooklyn DAC+, I listened for the improvements delivered when the Pro iDSD was powered by an Uptone Audio JS-2 linear PSU. Note that the nominal input specified for the Pro iDSD is 15V/4A, but the specs allow for external PSUs in the range: 9V/6.7A – 18V/3.35A. 

The JS-2 is a dual-rail linear power supply, built around a 100VA transformer core, with a current rating, at 12V, of >7A sustained, and >10A peak. As such, it more than meets the specifications. At just under $1000, it’s not cheap, but in my previous experience with the Brooklyn DAC+, the JS-2 elevated the performance of that DAC significantly. Could it do the same here?

Yes it could, and yes it did!

On Dances with Waves from Anouar Brahem’s evocative The Astounding Eyes of Rita album (ECM 24/44.1), the deep woody tone of the oud is much more visceral with the JS-2 powering the Pro iDSD. There is more energy and excitement in the music, and there is a sense of greater dynamics, from the soft percussive strokes of the bendir drum to the deep plucks of the bass.

Feature exploration: Filters

With the Pro iDSD now powered by a JS-2 rail, I next did an exploration of the rich feature set of the DAC. First up were the filters. iFi provides 5 of them: Bit-Perfect (BP), Bit-Perfect+ (BP+), Gibbs Transient Optimized (GTO), Apodizing (Apod), and Transient Aligned (TA). The first two are non-oversampling, while the other 3 oversample PCM to 705.6/768 kHz. 

iFi has published an entire paper on their GTO filter, so it clearly received a lot of design effort. I tried all 5 filters with a variety of PCM tracks, varying from Redbook to DXD, and ultimately settled on preferring GTO the most, with TA (Transient Aligned) a close second. Like all DACs with filter choices, I found the effect of these filter settings quite subtle, but over time a preference did emerge. I found GTO to strike the right balance between smoothness and resolution, with a neutral tonality. 

The nice part of course is that you can pick whichever suits your tastes. I’ve said before that I’m not a fan of tweaking filters endlessly, so for the rest of my listening, I set the filter to GTO and never changed it again.

DSD streams are not oversampled (unless DSD remastering is on - see below), so are always processed in bit perfect mode.

Feature exploration: DSD Remastering

DSD Remastering is a feature that targets customers who use software upsampling with powerful upstream hardware to upsample all sample rates to DSD512. iFi have elected to supply an FPGA-based implementation in the box, presumably to obviate the need for upstream processing. They explained the processing flow in the Other Questions section above.

Since I do not have a setup where I use upstream upsampling, I was not able to compare the efficacy of iFi’s FPGA implementation with say, HQPlayer. What I did do is evaluate the impact of iFi’s DSD Remastering on the sound quality.

Here again, I tried a variety of music, both PCM and DSD, with varying sample rates. On the whole, I mostly preferred DSD Remastering off. The main effect I heard was a taming of treble harshness, which could be helpful on more strident recordings. However, this came at the cost of a slight smoothing or rounding of transients. This effect was more pronounced going to DSD1024, so when I did use it, I found I preferred DSD512.

This is another feature whose benefits and value depend on the ears and preferences of the listener. It is a cool feature to have. For myself, I rarely used it.

Feature exploration: Solid-state vs. Tube/Tube+

Continuing on the theme of versatility, the Pro iDSD offers a choice of completely switchable output stages: either solid-state, Tube (using a GE 5670 tube stage) and the tubier Tube+.

While this was a fun afternoon of experimentation, I discovered that on the whole, I liked the Pro iDSD’s Solid-State stage far more than either of the tube stages. Don’t get me wrong, the Tube settings did imbue the sound with that characteristic tube warmth, but the solid state output stage on this DAC is so good, that it didn’t need any added color. The tube settings did seem to round off transients, which I didn’t prefer.

But here again, this is a great option to have, as I can see tube lovers having a different preference than mine. I love the choices this DAC provides!

Comparison: Pro iDSD/JS-2 with the Mytek Brooklyn DAC+/JS-2 

With the Pro iDSD now powered by a JS-2 rail, I was curious to see how it would compare with the Brooklyn DAC+, also powered with the same PSU.

Haydn’s Creation has dozens of recordings to choose from. This one by Andreas Spering and the Capella Augustina (Naxos DSD64) is one of my favorites, both for its performance and its sonics. 

Track 26 begins with a chorus singing Vollendet ist das grosse Werk. Both DACs rendered this beautifully, with each displaying different strengths. The Brooklyn was the more dynamic, going a bit deeper and fuller, while the Pro iDSD was the slightly more resolving of fine details.

Then the track continues to the hauntingly beautiful Zu dir, o Herr, blickt alles auf, sung by the three angels Gabriel (soprano), Uriel (tenor), and Raphael (bass). The Pro iDSD did a better job conveying the vocals, and the emotive singing. 

Honestly, just as with the stock PSU comparison, I would be hard pressed to pick one of these DACs over the other on sonics alone. 

Improvement from Ref-10 reference clock

Having validated that the Pro iDSD scales up with PSU quality, I turned my attention to another dimension of scalability - clocking. I doubt anyone will argue that clock quality matters profoundly in a DAC. It’s more about how high quality clock signals are delivered to the relevant circuits in the DAC.

For the vast majority of DACs, manufacturers will source oscillators of their choice, and implement their clocks internally in the DAC. Some high-end and pro manufacturers have provided inputs for a “word” clock, which is an external clock operating at, or multiples of, the 44.1 and 48 kHz sample rate families.

A second form of external clocking is called reference clocking, and this is what I will be testing. In this mode, the DAC’s clock design comprises an internal 10MHz reference clock, from which are generated (synthesized) the other frequencies needed across the device. In this design, the internal reference clock can be overridden by an external reference clock of higher quality.

Enter the Mutec Ref-10. I happen to already be using this extremely high-quality reference clock in my chain - to provide a reference clock signal to a SOtM tX-USBultra. Since the Ref 10 has 8 clock outputs, 6 of them at 75 ohms, I had plenty of available outputs to drive the Pro iDSD, which requires a 75 ohm BNC cable. Functionally, I had to set the Pro iDSD’s rear panel clock switch to the atomic clock position (atom icon) as shown:

Once the reference clock is attached to the BNC input, there is no indication at all on the display that an external clock has been detected. Ideally, that would have been a nice feature. As such, I just had to rely on my ears.

So what was the impact of attaching the Ref-10? One word - WOW. The addition of a reference clock of this quality transported this DAC into another class.

Listening to The Chain on Fleetwood Mac’s Rumours (24/96 from the DVD-Audio release), the addition of the reference clock is a profound improvement. The entire soundstage opens up and grows in 3-D space. Music and instruments snap into focus. It’s analogous to swapping in a sharper lens on a camera. There is more air around the instruments. The rattling of the, well, chain at the opening has more texture and realism. The drum beats are more meaty and realistic.

But let’s put this in perspective. This is a scenario that very few Pro iDSD owners will ever try. The Mutec Ref-10, at $3595 MSRP, exceeds the price of the Pro iDSD! So what’s the point of this comparison?

Scalability. I told you this was the theme of this article. In a purely technical sense, it’s exciting to learn how much better sound quality this DAC is capable of delivering. Secondly, this finding is relevant for those, like me, who are already using a reference clock in their chain - for DDCs (digital to digital converters) like the Mutec MC-3+ USB, for renderers and USB regenerators like the SOtM sMS-200ultra and tX-USBultra, and the like.

If you happen to be in that group, the $2495 investment in the Pro iDSD could yield a DAC solution that competes with much pricer DACs. Let’s look at the competition.

Comparison: Pro iDSD/Ref-10/JS-2 with Denafrips Terminator

Before I get into sonic impressions, let’s take a stab at what we’re comparing here from a price perspective. Since the JS-2 is a 2-rail PSU, we’ll value one rail at half MSRP, or $500. And let’s assume you’re using 2 outputs total of the Ref-10, so one output is $1800. By this accounting, the Pro iDSD/Ref-10/JS-2 combo is valued at $4800. That’s pretty close to the Terminator’s MSRP of $4450. So how do these configurations compare?

On Só danço samba from the iconic Getz/Gilberto album (Verve DSD64), these 2 DAC configurations exhibited different strengths. The Terminator impressed with its now-familiar textural density, deep bass, and muscular dynamics. The Pro iDSD/Ref-10/JS-2 combo was no slouch either. It conveyed more ambience, more fine detail in the instruments, especially the percussion, and better separation and imaging. To my ears, the Pro iDSD combo won the day, but this wasn’t a slam dunk by any measure. 

Again - it’s worth repeating that this isn’t a $2500 DAC trouncing a DAC twice its price. We all like giant killers, but this wasn’t it. This was two similarly priced DAC configurations exhibiting different strengths.

Improvement from SR-7 power supply

Ah - but we’re not done on our scaling journey. Stepping up another notch on the PSU axis is the Paul Hynes SR-7. In its tricked out form with double-regulated (DR) rails and silver DC cables, we’re talking about ~$1300 per rail for this supply. What does stepping up from the Pro iDSD/Ref-10/JS-2 combo to the Pro iDSD/Ref-10/SR-7 combo yield?

On the 4th movement, Allegro, of Jan Sibelius’ Symphony No. 4, Osmo Vänskä, Minnesota Orchestra (BIS, DSD64), the SR-7 brought a comprehensive improvement. Front and center is an increase in dynamics and low-end weight. The SR-7 opens up and expands the soundstage, with more air between instruments. At the same time, the sound is more refined, tonally more natural. Instruments are more dimensional and easier to tell apart.

Comparison: Pro iDSD/Ref-10/SR-7 combo with Ayre QX-8

At this point, I’d maxed out the Pro iDSD combo, with the best PSU at my disposal (SR-7) and the best clock at my disposal (Mutec Ref-10 reference clock). Price-wise, using the same accounting as before, this combo was worth $2500 (Pro iDSD), $1300 (one rail of an SR-7) and one clock output of a Ref-10 ($1800) for a grand total of $5600.

Well, what do you know, that’s right in the ballpark of the Ayre QX-8 ($5450 with USB and network).

Thomas Dausgaard’s recordings with the Swedish Chamber Orchestra on BIS are a true treasure. I have yet to find one I didn’t like. This recording of Tchaikovsky’s Symphony No. 6 “Pathétique” (BIS, 24/44.1 via Qobuz) is no exception. In the 3rd movement, Allegro molto vivace, the Pro iDSD combo puts in a really strong showing. The texture and detail of the massed strings is excellent. Individual instruments like the clarinet and the piccolo are extremely well rendered. Overall the Pro iDSD combo has slightly better dynamics and deeper bass.

The QX-8 shines in tonality. Instruments sound more natural, and there is a more relaxed and refined presentation than the Pro iDSD. One of the key strengths of the QX-8, from my previous evaluation, was its ability to portray a large and deep soundstage, and here too, it edges out the Pro iDSD combo.

All in all, though, these two combos were both exemplary. Each had its particular strengths, but I wouldn’t declare one unambiguously better. Strictly on my personal preference, I’d give the nod to the QX-8. But when you take into account that I already own a Ref-10 clock, and will very soon take delivery of an SR-7 PSU, the iFi Pro iDSD is by far the better buy. It’s a different equation based on where you start.

Comparison: Pro iDSD/Ref-10/SR-7 combo with Ayre QX-5 Twenty

As with most of my previous DAC evaluations, this is the moment of truth, when I (unfairly) compare the DAC combo under test with the almost $10k QX-5 Twenty.

I was recommended this version of Bach’s Brandenburg Concertos, Rinaldo Alessandrini, Concerto Italiano (Naïve, 16/44.1 via Qobuz)  by another CA’er, and it has now become my all-time favorite performance of this beloved collection. In the Menuet & Trio 4th movement of Concerto No. 1, the QX-5 allows the winds to soar in their full-throated glory. As it has with previous DACs, the QX-5 produces more of all the critical sonic attributes necessary - deeper and better articulated bass lines, smoother - almost creamier - tonality, better instrument separation, a bigger and more enveloping image, and instruments that are more fleshed out.

The lesson here is that while scalability is a desirable attribute in a DAC, it can only take you so far. In this case, while scaling the quality of the PSU and the clock yields impressive improvements, there are many other aspects of a DAC design that affect the sound quality of a DAC that we could not scale. So it should come as no surprise that a much more expensive DAC like the QX-5 Twenty still outperforms the scaled Pro iDSD.

Comparison: Pro iDSD/JS-2 combo with Schiit Yggdrasil with A2 upgrade

As is my habit with these evaluations, I like to try things out in systems other than my own, and with speakers instead of headphones. This is always interesting to validate findings in a broader context that is not system-specific.

So as with the other DACs in my evaluation, the Pro iDSD took a trip across town to my friend Eric's speaker-based setup, to compare with his Yggdrasil. The exciting news this time was that he had received his Yggy back from Schiit a few weeks earlier, with the Gen 5 USB and Analog 2 upgrades. This was the first time I’d be comparing a DAC with the upgraded Yggy, and it was exciting.

His setup comprises an Innuos Zenith SE server, driving a chain of SOtM tX-USBultra and dX-USB HD Ultra to the DAC via AES/EBU. Both the SOtM boxes are powered by independent DR (dual regulated) rails of a Paul Hynes SR-7 PSU. 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.



His analog chain comprises an Audio Research Reference 6 preamp, Hegel H30 power amp, and Magnepan 3.7i speakers, with dual Rhytmik F12g subwoofers. 

We picked the Pro iDSD/JS-2 combo to compare with the Yggy A2, since we had done the same combo (DAC/JS-2) with the Brooklyn DAC+. Eric had wisely refrained from telling me his opinions on the changes wrought by the A2 upgrade. Even though I don’t listen to the Yggy every day as he does, I knew something positive had happened when we started listening!

On Morning Bird from Sade’s Soldier of Love (Epic, 16/44.1) album, the Yggy A2 sounded much better than I remembered it. Bass extension and dynamics have always been its strong point, and there were prodigious amounts of both. In addition, the sound stage was detailed and airy, both more so than I remembered in the original. 

The Pro iDSD/JS-2 combo was no slouch either. It had similar dynamics but sounder a bit leaner in the bass. Imaging seemed smaller and more constricted. Tonally the sound  was smoother and very pleasing, but did not have quite the same fine detail as the Yggy A2.

Based on this comparison, I’d give the edge to the Yggy A2.

Enter the scalability feature of the Pro iDSD. Stepping up PSUs from the JS-2 to the SR-7 on the Pro iDSD changed the picture quite dramatically. The sound stage opened up and expanded, dynamics improved, and there was a wealth more of fine details. Whereas the Yggy A2 sounded better than the Pro iDSD/JS-2 combo, the Pro iDSD/SR-7 leapfrogged both to sound the best.

The key takeaways from these eye-opening comparisons are that:

1. The Yggy with the A2 upgrade is a heck of a value, provided you don’t need more than its feature set.
2. The Pro iDSD, in addition to its far more extensive array of features, has an unparalleled capability to scale up to very high levels of SQ - albeit at a price.

Both approaches have value and appeal - it just depends on your requirements.

Summary

This has been a long and challenging evaluation, because this is really the evaluation of multiple DAC, PSU, and clock combos. The iFi Pro iDSD is one of the most versatile and feature-laden DACs I have had in my system. 

1. Considered as a standalone DAC, with no upgrades, the Pro iDSD is a standout value at its price point. This would be true based on its impressive feature set alone, but once you add in its outstanding sonics, I’m hard pressed to think of a DAC that can match it at $2495.
2. But wait, there’s more. The Pro iDSD’s ability to scale up PSUs and clocks gives a buyer an upgrade path to better sound quality without needing to trade in anything.
   1. The Pro iDSD was able to accept PSUs of increasing quality all the way up to a Paul Hynes SR-7, and produce commensurate increases in sound quality.
   2. Further, the addition of a reference clock like the Mutec Ref-10 yielded another dramatic gain in sound quality, albeit at an increased price.
3. While these upgrades represent an increase in overall cost, the resulting combos matched up to comparable standalone DACs. The Pro iDSD/SR-7/Ref-10 combo stood head-to-head with the Ayre QX-8, which is no mean feat, given how outstanding the latter is.

I think iFi have a real hit on their hands. The Pro iDSD appeals to many distinct audio constituencies. To owners of an entry-level DAC, the Pro iDSD is a formidable upgrade, both in features and sound quality. To owners of a mid-price DAC, the Pro iDSD combined with a high-quality PSU like the JS-2 would be a compelling upgrade. To tinkerers and tweakers like me, who already have high-quality PSUs and a reference clock, the Pro iDSD provides an extremely economical way to achieve a level of sound quality associated with $5k+ DACs.

Bravo, iFi!

Product Informtion:

• Product - iFi Pro iDSD ($2,499)
• Product Page - iFi Pro iDSD
• PDFs - Pro iDSD User's Manual

Associated Equipment

Here is a diagram of how the Pro iDSD integrated into my system:

Music Server:                 Innuos Zenith Mk II SE
Headphone Amplifier:   Cavalli Liquid Gold
Headphones:                  Sennheiser HD800 (Super DuPont Mod), Audioquest Nighthawk
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 6 AWG 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 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 loom of Cardas Clear cables to allow identical cables to be used for all the DACs under test!