The Background
I originally set out to review the Geshelli Dayzee, and I still intend to, but before I got the chance, I ended up experiencing the Chord Mojo 2. The Mojo 2 really got me thinking about what is possible or desirable in a DAC. My ever-vigilant upgrade gremlin saw the opportunity and made me purchase a Chord Qutest. This left me with 3 DACs in roughly the same price range with very different attributes – the perfect setup for a shootout!
Once I decided to write something up on this for public consumption, it occurred to me that I had seen questions on the topic of the advantages of the Geshelli Dayzee and its dual mono AK4499EX chips over other DACs with the AKM AK4499EX. Unfortunately, I did not have a J2S handy for the comparison, but I do have a Topping E70 Velvet, which I know is different, but it is in a similar price category to the J2S and has a single AK4499EX chip. So, the E70 Velvet became a wildcard 4th entrant in this comparison to give some imperfect insight into the relative standing of a DAC with a single AK4499EX chip versus competition in the next pricing tier up.
Time for some obligatory throat-clearing. In our world, diminishing returns are an accepted reality, and they absolutely do not stop us on our quest for sonic bliss. There is not a bad DAC in this lineup. All comparisons are intended to explore the differences between these products, so individual differences will be pulled out and dissected in far greater detail than might be reasonable or even particularly noticeable in listening to any of these devices casually or even more intently on their own outside of the setting of an A-B-C-D comparison. Also, this comparison is based purely on sound quality, with no consideration of features, appearance, etc. These are all important, and they definitely can affect the experience in different ways, but they will depend very heavily on an individual’s use case. With all of this out of the way –
Let’s meet the contenders!
Denafrips Ares 12th + Iris 12th
For many readers, the Denafrips Ares will need no introduction, but just in case... The Ares 12th (for 12th anniversary) is the successor to the wildly popular Ares II. It is an R-2R resistor ladder-based DAC. The Ares 12th uses a custom USB FIFO interface and FPGA to handle oversampling of incoming PCM to sampling rates in an integer multiple of the source to a max of 1411.2 or 1536 Hz. DSD is not upsampled or converted to PCM but is handled through a separate hardware filtering system specifically included for native DSD support. The Ares 12th also sports a high-quality internal linear power supply and carefully matched resistors to achieve relatively high measured performance, particularly for an R-2R device. The Ares 12th was very quickly superseded by the Ares 12th-I, with the 12th being Denafrips’ primary lower-cost DAC for less than 1 year. Comparisons between the two are available online, but it appears that most of the changes dealt more with quality of life and form factor than sonic performance, so impressions here should be very close to the 12th-I.
The Ares 12th features a nominally non-oversampling (NOS) mode, however, there has been some controversy about whether this feature represents true non-oversampling. This controversy started with the Ares II, and the 12th purported to fix this issue, however, testing from GoldenSound would appear to indicate that, while the results are much more similar to what would be expected by NOS, there are signs that the actions being taken by the DAC do not correspond to what most people regard as NOS. I mention all of this because NOS (particularly the update to “real” NOS with the 12th version) was a major selling point of this DAC and because readers will wonder if it was used in this comparison. I have compared the sound of NOS and OS mode on the DAC with Roon delivering native bitrate files, and to my ears the OS mode is a hands-down winner in sound quality across the board, so the Ares 12th was in OS mode for the entirety of this comparison. The firmware was version 1.8 (the latest available I believe) for this test.
The Ares II received a huge amount of fanfare as a game-changing DAC in the roughly $1,000 price range. When I was looking into picking it up, the Ares 12th was just coming onto the market and appeared to have all of the benefits of the Ares II with a few significant upgrades, so I opted for the new model. It sounded great right out of the box, and it was definitely different from the Matrix Audio X-Sabre Pro MQA that I was using previously. I could hear moments of greatness listening to it, but it definitely didn’t feel as Earth-shatteringly amazing as reviews of the Ares II made it out to be (part of that may have been down to the wonky firmware version that shipped on it initially, and which is no longer even available for download). I was really excited to hear for myself what all the R-2R buzz was about, and I had a distinct feeling I was missing something. That’s when I found the Iris 12th.
The Iris 12th is a DDC, and it appears purpose-built to accompany the Ares 12th in appearance, form factor, and features. I am definitely in the “bits are bits” camp when it comes to digital audio, so I was extremely suspicious of the concept of a DDC, but once again GoldenSound provided the testing that showed that there was a possibility of a measurable difference in actual analog audio output of the chain when adding a DDC. There are multiple possible reasons for this, including more precise timing with the higher quality clock available in the Iris 12th and optical decoupling of electrical signal to prevent noise from bleeding into the DAC from the ground wire of the source being the most compelling to me. One of the other standout features of the Ares 12th over the II was the addition of an I2S HDMI port, which is the preferred connection for the Iris 12th.
I decided to take the gamble on adding the Iris 12th to my system, and it made an enormous difference in refinement and delivering the “magic” I was expecting from the Ares 12th. The combination of the two devices delivered what I was expecting from the Ares 12th based on the hype it was getting. This was reconfirmed when I was moving the Iris 12th and accidentally blew the fuse in it. During the week or so I was waiting for a replacement fuse to arrive, I got to experience its absence after having used it for the better part of a year, and the difference was greater than I remembered (of note, this was still prior to upgrading the firmware on the Ares 12th).
The current version of the Ares retails right around $1,000 and the Iris around $600, though they fluctuate a little because they are sold in Singapore dollars, so exchange rates are a factor. These approximate prices are in USD.
From this point on in this comparison, when I refer to the Ares, I mean the Ares 12th + Iris 12th combination.
Chord Qutest
Many readers will know more about the Qutest than me. I have heard of Chord for many years, but I had never tried any of their products until I stumbled upon a Rob Watts video on YouTube. I had heard him interviewed on a couple of occasions, but in this case, he was giving a presentation at CanJam NYC 2024. (Major shout out to MidFi Guy for putting this on his YouTube channel!) Rob was making a huge amount of sense to me in the way he was describing his process of subjective and objective device testing and refinement. There was something about his presentation that threw me, though. He seemed to have a disdain for the vast majority of the DAC industry, which led me to think that he is either a totally delusional narcissist or perhaps he is onto something. Later it occurred to me that maybe my knee-jerk reaction was a little extreme, and he is an engineer, not a salesman, and he is very passionate about his work... In any case, I was curious about his products, and I started to look for reviews, and virtually all of the people I trust seemed to be over the moon about most if not all of the Chord DACs. I picked up a Mojo 2 and immediately realized I was dealing with something different. This very quickly led me to the Qutest, which felt more “on par” with the level of equipment I have been using.
The Watts philosophy is almost the inverse of what I have learned about what goes into making an excellent DAC. The traditional wisdom, as I understand it, is that converting digital to analog is a relatively straightforward process and has basically been obviated by modern technology, after all, the Apple Dongle has a tiny DAC chip that measures almost perfectly… Oh! Sorry, wrong forum
It is further asserted that all that really matters in terms of sound quality in a DAC is the analog amplification and output stage. Sure, there are different chips (Burr Brown, AKM, Sabre, Cirrus, etc.), but most of them, it is said, sound almost identical, and it is the linear power supplies and analog amplification circuits that make the difference between an excellent DAC and a mediocre one.
Chord and Rob Watts, on the other hand, believe that the digital side of the DAC is where the magic happens. The following explanation is ludicrously simple relative to the info available on YouTube, HeadFi, etc., but for the uninitiated, I will attempt to set the stage. Using much higher performing custom-designed FPGA chips enables extreme oversampling of incoming data to more accurately reconstruct the timing of transients. This is combined with careful attention to eliminating or attenuating any sources of noise floor modulation that may vary with the signal. The result is supposed to be extremely precise timing of transients in the music, allowing for a more natural experience of soundstage and timbre. There are also other effects, such as extremely good objective filtering performance with a near vertical drop in signal response around 22 kHz. I hope I have done justice to the basic philosophy of the Chord DACs here, and please feel free to point out anything I missed in the thread below, but if nothing else, this shows roughly what I’m thinking going into this.
One other interesting note with Chord’s devices: they tend to be some of the oldest designs that remain in production at this time. Purchasing the Qutest, a digital electronic product released in 2018, in 2024 feels very odd in a climate where devices are constantly one-upping each other (see above - my Ares 12th was current for less than 1 year). It is not the oldest Chord product still in production. Rob has said (I’m paraphrasing) that this is because competitors are not innovating and are making incremental improvements on a fundamentally flawed design, so there is no pressure to innovate on a faster timeline at various price points. Considering the awards that Chord products rack up year after year and the ability to continually sell units, it seems like he might have a point.
The Qutest is a standalone DAC with dual BNC coaxial (can be used together with certain specialized products to double bandwidth), optical, and USB inputs. The USB input is galvanically isolated to eliminate electrical noise from the source traveling into the DAC. Power is handled by a wall wart delivering 5V DC power via a micro-USB connection. Its only output is via unbalanced RCA connectors. It features 4 filtering settings, Neutral Incisive and Warm – each with the option for adding high frequency roll-off. I did my listening with Neutral Incisive, no roll-off. I believe that this represents the full capability of the device and should allow any special abilities it may possess to shine through. Output voltage is selectable and can be set at 1V, 2V, or 3V at 0 dB. I did my listening at 2V, as this was closest in perceived volume to my other DACs’ RCA outputs. I did not notice a significant difference in sound quality between the settings, just a slight difference in the volume knob position on my amp.
The Qutest currently retails for approximately $1,600.
Geshelli Dayzee
This is Geshelli Labs’ flagship DAC, designed to fit into high-end hifi systems and expand the Geshelli DAC range beyond the extremely popular J2S – which has been heralded as an excellent value for a first serious audiophile DAC as well as a highly aesthetically and sonically customizable device. The Dayzee features 2 of the very highly regarded AKM AK4499EX DAC chips - one for each stereo channel. The only significant audio customization offered is in the form of op amps (a very popular aspect of the J2S, where S stands for socketed, that has carried over to the Dayzee). In the default configuration, there are 8 op amps, 4 per channel, that can be swapped. Unlike with the J2S, these op amps are not obviously associated with specific outputs and cannot be changed individually. All op amps must be swapped if the user desires a different option.
The Dayzee ships by default with OPA1655 op amps. There is an option to swap in the Sparkos SS2590 – the professional bigger brother of the very popular SS3601 discrete op amp. The SS2590 is a larger op amp on a 6-pin professional package with a significantly larger footprint than the DIP8 op amps included by default. In my very early production unit, there was a small hiccup in communication and sockets for the larger pro op amps were not included. This led to an opportunity to break out the soldering iron and attach them to the board myself (more fun than inconvenience in my mind). I only mention this in the service of full transparency, in the extremely unlikely event that this would color the sound (or avoid coloring it) in some way relative to plugging them into a socket. The lovely folks at Geshelli have assured me that all future Dayzee units will have sockets to drop in the pro op amp upgrade regardless of the option selected at purchase.
The Dayzee has dual XLR balanced outputs, as well as dual RCA unbalanced connections. These are all functional simultaneously. Its 6 inputs include USB (galvanically isolated to prevent electrical noise from sources entering the DAC), AES, and 2 each of optical and coaxial (RCA connector). The interface is extremely simple. There is no remote control, filter selection, volume attenuation, etc. All you have is a row of buttons, one of which is a power switch. The others each correspond to a specific input. Notably, DSD is only supported via the USB input, with the others being apparently limited to 192/24 PCM. Power comes from what appears to be a fairly standard wall wart rated for just over 2A of 12V DC.
There are a variety of cosmetic customization options available, and the price of the Dayzee can vary significantly depending on which ones are selected. The base price starts around $1,300 and 8 of the Sparkos SS2590 will run you around $480, though you can get a discount if you order them directly from Sparkos and sign up for their email list. This brings the package up to nearly $1,800 and into a whole new realm of competition for Geshelli.
Topping E70 Velvet
The Topping E70 Velvet (E70V) is the AKM version of the well-regarded E70 that used an ESS Sabre chip. The E70V has one of the AK4499EX chips to handle both left and right channel duties. The E70V, like most of Topping’s offerings, boasts extremely low measures of distortion and noise, delivering measured performance much higher than expected for its price. With the Velvet version of the E70, a level of smoothness (presumably velvety smoothness) was to be added by changing to the AK4499EX without compromising the measured performance. In fact, the measured performance actually improved slightly.
I have been enjoying the E70V for some time now, mostly for near-field speakers hooked up to my computer. I had moved it for a while to my headphone setup, but as I started to get serious about head-fi in the past year, it was relegated back to the computer in favor of the competitors mentioned here. Before this comparison, it had been several months since I listened to it with headphones.
For inputs, the E70V sports coaxial, optical, USB, and Bluetooth. Output is available fully balanced via XLR out or unbalanced over RCA. It also has a wide variety of features, including selectable filters, remote control, digital domain volume attenuation allowing for function as a preamp, adjustable voltage output on XLR (4V or 5V) and RCA (2V or 2.5V). The 4V and 2V output levels were used in this comparison. Power is handled by an internal SMPS, connected by a standard 3 prong cord.
One interesting aspect of the AK4499EX chip is that it really is a set of two chips, one analog and one digital. This allows for better isolation of the analog stage from the digital conversion, an inherently noisy process. I don’t want this to be confused with the Dayzee, which has two of the AK4499EX, with each of these consisting of two chips (analog and digital). This dual-chip configuration is part of what raises the price on devices featuring the AK4499EX over their counterparts. There are significant differences between the E70V and the Geshelli J2S (also available with a single AK4499EX in its top configuration, which is notably pricier), but this is the closest I have available for comparison.
One final fact about the E70V will be especially exciting to the value-oriented DSD enthusiast. The E70V, when run in pure DAC mode (with volume attenuation disabled), is able to bypass the delta-sigma conversion process completely, allowing native handling of DSD. When volume control is enabled, the DSD signal is converted to PCM and fed through the same D->A conversion pathway as PCM sources. The native handling of DSD is purported to give the AK4499EX an advantage in DSD playback, but only when correctly implemented. Some devices with the AK4499EX do not take advantage of the ability to bypass the PCM conversion step. The E70V was running in pure DAC mode for this comparison to take advantage of this feature. Tom Gibbs has a fantastic piece detailing this aspect of the E70V on the PS Audio website - https://www.psaudio.com/blogs/coppe...tive-dsd-boom-with-the-topping-e70-velvet-dac
I believe that the E70V is one of the least expensive implementations of the AKM flagship AK4499EX at $450.
The Setup
I am not a professional reviewer, and this test was not intended to be scientific, but I did try to control as many of the variables as feasible. I do believe that the power situation in my home is a bit prone to noise and voltage fluctuations, if the frequent, sudden changes in brightness of the lights are any indication. To give each device its best chance to shine, I used a Pine Tree Audio combined AC / DC power filter in front of a Furman AC filtering power strip. Two of the DACs (Ares and E70V) have fully internal power supplies, so this is all the help they’re getting. The two with wall warts got a little extra attention in the form of iFi iPower X adapters in their respective voltages in place of their included wall warts.
I listened to 6 tracks with which I’m quite familiar and took notes. I listened to a track on all 4 DACs all the way through prior to moving to the next track. I kept the order of the DAC rotation the same, but I started with the next DAC in the rotation on each track, so each one got the benefit of being the first to play at least one track.
All listening was done on my ZMF Atrium with universe suede pads and stock mesh connected by ZMF 2K Copper cables (1/4 inch for the single ended amp and 4 pin XLR for the balanced one).
The first 3 tracks were tested on the Woo Audio WA2 amplifier – an output transformer-less, single-ended triode tube amp. Tubes used were NOS Mullard-made E188CC drivers, NOS Chatham 5998 power tubes, and NOS Mullard-made EZ80 rectifiers. All of these tubes are in good condition and have between 100 and 300 hours play time depending on the pair. This amp has also been modified with the interstage coupling caps having been replaced with Mundorf Supreme EVO Silvergold (no oil). Each DAC was connected to the WA2 by a set of Kimber Kables Hero RCA cables.
The last 3 tracks were tested on the Violectric V222 amplifier – a fully-balanced solid-state amp. All 3 user-accessible op amps have been replaced with Sparkos SS3602 discrete op amps. As this amp uses a fully balanced topology, and as I understand it, this is the preferred way to connect devices to it, I used World’s Best Cables Mogami star quad XLR cables to connect the 3 balanced DACs, and I used the Kimber Kables RCA from the other test to hook up the single-ended-only Qutest.
Volume was not scientifically matched. I just sought to start slightly low and increase to the loudest comfortable listening level as quickly as possible with each track (I believe I tend to listen somewhere on the threshold of moderate to loud, but I don’t have the setup to measure SPL to be sure).
4 tracks were streamed from Qobuz and 2 from my files. Rips of the 2 tracks from my own physical media collection were done with great care to get bit perfect copies – one PCM from Red Book CD and one DSD from SACD.
Playback was handled by Roon. All tracks were played in their native format with no DSP (fixed volume in Roon, purple “lossless” label present). My Roon Rock NUC streamed all tracks over WiFi to a Raspberry Pi 4 running RoPieee. The Pi was connected to all 4 DACs simultaneously by USB using generic cords provided with the devices or similar equivalent. Note that all of the options have some sort of galvanic isolation except the E70V, so I used a Topping HS02 between the E70V and the Pi to galvanically isolate, so noise from this source should not be problematic. I assume most people looking at DACs in this level of depth will have an understanding of the importance of isolating from an electrically noisy source, so I didn’t want to handicap the E70V, as most people on HeadFi will probably use a streamer built for low noise or an isolating strategy if using a DAC that does not have galvanic isolation.
The Comparison - Listening Notes (presented in order of listening)
Round 1 – Woo Audio WA2 SET OTL Tube Amp
Saturnine – GoGo Penguin – Qobuz 48 kHz 24-bit
Dayzee – Synth at the beginning of the track is clean and well-defined, and the low end is particularly powerful. Moving to string bass, higher notes start to have a slightly harsh character – detailed, but slightly grainy or over-sharpened. Timbre is good, nowhere is it particularly problematic in this track. Piano has excellent note weight – not overbearing or excessively emphasized, but very enjoyably engaging. Left-right sound stage is presented well, but not much depth is noted – everything feels intimate. Tonal balance is good, but it feels like there is a very slight bass boost.
Ares – The reverb and ambiance created by the opening synth are immediately arresting. Bass is adequate, not exaggerated. The tonal quality (as opposed to transient) of the snare drum is very noticeable. Timbre of the piano suffers relative to Dayzee – leading edge feels smeared, and it sounds like it could be a piano or synthesizer. There is a real sense of space right from the beginning and instruments are spread evenly in the soundstage with no “blobs” emerging. There is a modest sense of depth – particularly noted on snare drum. There is a slight feeling of an element of height elevation with piano. There are no harsh or grainy tones. Overall, a pleasing balance of note weight and air.
E70V – Movement of the synth in the beginning of the track around the soundstage is instantly apparent. This is not something I had appreciated with the first 2 DACs. The standup bass has a very noticeable sense of texture, but it feels slightly exaggerated – not fatiguing, but doesn’t seem as natural as the others. Good sense of timing and rhythm – feels faster than the first 2. Feels more detailed overall than the first 2, unclear if exaggerated overall or appropriate to the recorded material. No significant depth is appreciated at all. Tonal elements of percussion are almost completely absent here.
Qutest – Air and sense of ambiance are very noticeable right away, similar to Ares, maybe slightly more. Movement of synth is appreciated again here, but not as dramatically as with the E70V. Some texture is present on upright bass – not as much as with Dayzee or E70V – no harshness nor exaggeration at all. Timbre of the piano and percussion are absolutely convincing. Note weight is adequate, but feels lighter than the others, unclear which might be closer to the intent of the recorded material. Detail is highest of all DACs tested here – no exaggeration, harshness, nor etching. Left-right spacing is good. Depth is still modest on this track, but more than with any of the other DACs
Paper Tiger – Beck – MoFi CD 44.1 kHz 16-bit
Ares – Intro vocals feel very intimate. There is a great deal of air and separation around all instruments. There are subtle details present in the electric guitar. Handles busy passages well throughout the frequency spectrum – maintaining good sense of air and detail when things are complex. Reverb is pronounced. There are harsh frequencies in this recording to my ears, and they are tamed adequately by the Ares. Generally pleasing and laid back but leaves something to be desired in terms of slam and macro-dynamics.
E70V – Reverb feels very sharply detailed – sound does not feel like it diffuses through space. There is significant detail in this presentation, sounding slightly dry. “S” sounds in vocals are dangerously close to excessively sibilant, but not quite crossing the line. Soundstage is wide, but there is absolutely no depth – it has the sense of being on a cellophane film cutting through my head from ear to ear – not that it is harsh, but that there is no sense of depth whatsoever beyond the line between my ears. There is a buzzing character to the timbre on string instruments that I would attribute to overly etched detail. Still not particularly fatiguing. Presents a generally enjoyable experience in spite of some of the shortcomings that make it slightly less natural.
Qutest – Reverb on the opening vocals is subtle, detailed, sounds like a natural reaction to a space in which they are being sung. The bass guitar riff is a great deal easier to follow with all articulation present – this is a significant departure from prior 2 DACs. Sibilance in the recording is clearly noted but feels as though it has somehow been rendered inoffensive. Sounds of guitar strings responding to the individual frets as the musician’s fingers slide across are present in a way I don’t recall having heard before in this track. Complex passages feel effortless. Timbre of cymbals is particularly convincing. Micro-dynamics are excellent. Macro-dynamics are fine, but not particularly stunning.
Dayzee – Less sense of space and reverb on intro than the other DACs. There is a pleasing enhancement of note weight present in vocals that was not there for the other DACs. Detail is great, not forced. Sense of depth is minimal, but more than E70V. Imaging and left-right separation of instruments, as well as ability to pick them out of the mix are very good. Timbre of the percussion is notably good, but not quite on the level of the Qutest. Feels as though there is a theme throughout this track of note weight / tonality being enhanced over timbre and general technicality.
Creepin’ In – Norah Jones w/ Dolly Parton – Analogue Productions DSD (SACD)
E70V – Nice sense of rhythm and timing throughout – toe-tapping goodness. Detail is good overall. Maybe a tiny bit of excessive grain applied to Norah’s voice. Dolly’s voice sounds perfectly natural here. The bass line can be followed, and intensity is appropriate, but it is not particularly well-articulated, and there is minimal timbre information in the lower instruments. Melody creates an emotional response that is noteworthy, particularly given what I would have expected from this DAC. There is a sense of spatial depth that emerges in this track, particularly in the vocal harmonies.
Qutest – Norah’s vocals are presented with a striking level of detail but with no sense of artificiality or harshness at all. Similarly, Dolly’s voice is more textured and never harsh. Both sets of vocals feel like masterful blends of tonality and texture, presented naturally. The bass guitar is noticeably better articulated than with E70V. Generally, the instruments, like the vocals, have a tremendous balance of tonal information and realistic timbre. It is genuinely difficult to take notes here, as this is such an enjoyable listening experience.
Dayzee – There is a bit less air / separation of instruments relative to the 2 previous DACs in this track. Similarly to the 2 prior tracks, the Dayzee delivers a slightly elevated bass register. Detail is lost in the lower instruments here, similar to E70V. Note weight is enhanced relative to the other 2 DACs. Any harshness detected in Norah’s voice is totally smoothed out and pleasant here. Dolly, on the other hand, sounds a little bit distant and muffled. Timbre of percussion is good but fades in busier passages. All of the vocals have enhanced tonality – bringing some of the background vocals forward. There is some perceptible depth, but it is modest.
Ares – Even with raising the volume, the entire presentation sounds recessed, as though there is a sense of depth being applied, but to every element simultaneously. Backing vocals sound as though they have been pulled forward in space relative to the other DACs. There is a sense of things moving around in space as the track gets busier in a very odd and disorienting way that is difficult to describe. There is some very interesting detail in the tonality of the drums that I did not notice on other listens. Generally, the spatial presentation is quite distracting and unnatural. The tonality is pleasant, but I cannot get into the song at all. I did not do a second critical listen, but I tested a minute of the track using resampling in Roon to 192/24 PCM and the performance was much better – in line with the quality level that would typically be expected for this DAC.
Round 2 – Violectric V222 Fully Balanced Solid-State Amplifier
The Hardest Cut – Spoon – Qobuz 96 kHz 24-bit
Qutest – The clapping starting early in the track and continuing behind the drums is something I had not differentiated as a specific element of the percussion until listening on this setup. Very fine details of 3 distinct guitars are clearly noticeable and can be easily identified even as all are styling simultaneously. A richness of vocal harmony in the pre-chorus is particularly stunning. Staging is intimate overall, but during the bridge and solo, space expands out before snapping back in in a way that feels intentional in the production of the track. Excellent balance of musicality and technicality.
Dayzee – Clapping at the beginning of the track is difficult to differentiate as the track builds intensity. Good detail is present in the guitars, and the note weight and tonality are more impactful relative to Qutest. Vocal harmonies in the pre-chorus collapse into a single tone, with a loss of distinct detail between the voices. This track is generally intimate, and this is especially apparent here. No depth is noted at all. There is no harshness, etching, or excessive detail. Though not particularly technical, this is a very musical, enjoyable listen.
Ares – Clapping at the beginning of the track stays clear as other percussion picks up. There is a dramatic sense of air and reverb on the clapping, creating a more spacious feel than the 2 preceding DACs. Depth and layering are also very evident in the vocals. The guitars are all detailed and can be differentiated, only slightly less technical than the Qutest. There is a particular sense of groove to the bass line on this listen that is well-articulated and fun to follow along with. The tonal richness is not quite as high as the Dayzee, but there is notably more detail. Qutest is more detailed than Ares and has better timbre, but Ares has a larger overall sense of space on this track.
E70V – Immediately noticeable increased intensity in the bass. Though elevated relative to the others, bass is driving and pleasant. Clapping is clear and mostly preserved. Interestingly, the reverb effect on the clapping comes through clearly, and yet there is absolutely no sense of perceptible space or depth generated by it – an odd effect. The level of fine detail on the guitars is surprisingly high and goes toe to toe with the Qutest. The overall presentation is tonally rich and musical, but spatially flat.
The Unclouded Day – Randy Travis – Qobuz 44.1 kHz 16-bit
Dayzee – Immediately musical and tonally rich. This enhancement of note weight is welcome in this somewhat bright and very busy production. Timbre is very good. There is a bit of perceivable depth, particularly in the layering of the vocals. There is good air and space around the many instruments playing, and it feels like they can be explored without any of them being pushed forward excessively. There is a slight enhancement to the lower registers that is noticeable and has a balancing effect on this track. Great listen overall, technically and tonally.
Ares – Relative to the Dayzee, it is immediately apparent that the timbre is more natural on the instruments. The bass drum feels somewhat less impactful. The upright bass still has a sense of driving rhythm and toe-tap-ability. Randy’s voice relative to the Dayzee feels lighter and airier – slightly ethereal. Individual voices in the vocal harmonies are a bit difficult to pick out, but they feel cohesive and musical in the way they combine.
E70V – Compared to the first 2 DACs, there is a sense of dryness, but it is subtle, and I don’t know if it would be noticeable if not for directly following the other 2. Generally, still musical and engaging tonally. Soundstage: flat as a pancake. The presentation is intimate, which is surprising after the other DACs. The vocal harmonies collapse together, and voices cannot be differentiated. The instruments feel detailed, but they do not have space nor air. It seems like etching and excessive sharpening to create the perception of detail destroys some real detail here. I suspect this is because of the brightness of the mix – I’m not sure what year it was recorded, but it strikes me that the producer had the 80’s sensibility that he had never heard a mix with too much treble. This is not good for synergy with the E70V. It is not terrible, but this is a surprisingly striking weak point for the E70V.
Qutest – Timbre and instrument spacing in this track the best of the 4 by far. There is plenty of room throughout the entire track to explore the details of each instrument or mentally zoom out and appreciate the natural and cohesive overall production. The high treble is still present, but similar to the sibilant S sounds on the Beck track, the Qutest does its trick here where it honestly presents the information but somehow keeps it from being offensive. The note weight in Randy’s voice is lowest with Qutest of the 4, but the texture and timbre are totally natural and effortless. Though this is my 4th listen in a row, it is clearly the most emotionally arresting (this is a song that has affected me emotionally since I was a child). I actually feel a bit choked up while listening, which I was not expecting.
Chocolate Chip Trip – Tool – Qobuz 96 kHz 24-bit
Ares – Reverb is in full effect throughout the track. Though this track is almost exclusively percussion, there is a significant and noticeable amount of tonal richness. Lots of air present around the various elements. The left-right soundstage is smoothly rendered and gapless. Overall presentation is coherent and timbre, particularly on the acoustic drums, is very good. Bass drum has good slam without feeling excessively enhanced or washed out. There is noticeable depth in the center of the soundstage where multiple elements overlap.
E70V – Tone is not as pronounced with the E70V, and there is less air around instruments. There is an enhancement of texture, and the buzzer-like elements of the electronic instruments are brought to foreground. Depth is modest, but this presentation is not totally flat like some others with this DAC. Timbre is decent on acoustic drums, but not as convincing as Ares. Left-right tracking is smooth with no blobs, but the overall soundstage is more intimate.
Qutest – For the first time, after having used this as a test track since shortly after it was released, I notice the slight out of phase vibrato (if that’s the right term) between two of the tones in the beginning – similar to what you get when you are tuning a guitar and are almost there, but not quite. The timbre and tonality are balanced and very natural. There is a very large amount of fine texture, but it does not feel etched, enhanced, or harsh. The lateral localization on this listen is the best I have ever heard from any setup on this track (not that I have ever had any that are much higher end than these) and the ability to pick the elements apart is on a level I have not experienced. Technicality is through the roof. The low end of the frequency spectrum feels thin relative to the prior 2 DACs - though I wonder if this is just absence of artificial enhancement, given the exceptional technical performance on this track. There is still significant slam on deliberate hard drum strikes. It is difficult to overstate how blown away I am by the detail and technical performance on this track. This track can make almost any DAC sound pretty awesome, but the Qutest really feels like it was taken to another level on this one.
Dayzee – Note weight and tone are notably enhanced here relative to all of the others. There is less air around the instruments, but there is some noteworthy depth present in the center of the soundstage and great, even left-right separation. Timbre is less convincing than all of the other DACs. There is a washing-out effect on tone in the acoustic drums from the low-mid register and down. Though it is not a particularly technically impressive performance relative to the other DACs here, it is extremely fun. The slam is off the charts, and it is not hard to imagine why someone might prefer the enhanced macro-dynamics and associated giggle factor over technicality or even optimal tonality. There is still a high level of technical proficiency here, but the slight bass boost buys you a surprising amount of ferocity at the cost of a bit of precision.
Conclusions aka TL;DR
This project has taken longer than I expected, particularly finding time for the writing, so I have had a lot of time to think about it. I have decided to break this conclusion into 2 parts: first, the general characterization of the DACs in relation to one another; second, a collection of individual impressions and takeaways that I think might be valuable or that surprised me in this process. This is also a good place to reiterate that all of these DACs are competent in all settings, and strengths and weaknesses are being highlighted to differentiate between 4 very reasonable options.
The Synthesis
At the very real risk of being painfully cliché, I feel like the most coherent analogy I can craft to compare these 4 DACs is a visual one – Imagining the performances of each of these DACs as artists’ renderings and the recordings themselves being the actual scenes being represented. I think the general differences can be likened to a choice of artistic medium and style.
The Qutest is an extremely realistic oil painting. The presentation is impeccably detailed, life-like, and without excessive interpretation. In general, it seeks to capture the essence of the scene, almost like a photo. The technical presentation is absolutely stunning, and prior to experiencing the Qutest (and the Mojo 2), I would have thought that this description would be incompatible with musical enjoyment – at least for me. I have been a longtime lover of vinyl, and for the past 5 years I have also been enthralled with tubes, so I have a reflex to equate high levels of detail with a lack of soul and a bright, edgy, harsh experience – like I had with the Matrix X-Sabre Pro. The Qutest has completely disabused me of this notion, and I am simply in love with the sound of it. I cannot believe that there has not been more of an effort by competitors to follow Watts’ approach and eschew the received wisdom of enhancing the output stage while relying on an off-the-shelf chip. To the extent this is happening, it appears to be in R-2R DACs, but Chord is proving that abandoning microchips is not the only way to get to a different sound. I think anyone with a level of audiophile sensibilities, on hearing the Qutest or Mojo 2, would have to acknowledge that there is something different going on. I believe it is a line of inquiry that deserves to be fully explored.
The Dayzee comes across a bit more stylized. To me, it feels like it is rendered in bold, vivid, primary color acrylics. There is a sacrificing of fine detail in the service of tonal richness and impact. I think of opaque, striking reds, greens, blues, and yellows with thick black and white lines separating them, or perhaps the colors abutting one another, but with minimal if any blending. This DAC is competent in timbre and spacing, but unremarkable at its price point relative to the competitors here. Depth and texture may be diminished, but there is a striking sense of tonality that can enhance the emotion being communicated in a melody regardless of valence. By avoiding adding any over-etched enhancement of detail or harshness, listeners can be content to relax and let the Dayzee put the melodies and harmonies front and center for accessible enjoyment. The most fun features of songs with intense bass and slam will be even more thrilling here. There is enough detail and technicality that it will not feel like anything is being missed unless it is specifically compared to one of a handful of more technical DACs.
The Ares 12th + Iris 12th combination is a realistic rendering done with pastels. There is a sense of realism and technical detail, but there is also an enhancement of air and a generally ethereal sense being imparted to the music. Tonal density is variable – most of the time it is opaque and deep, but in certain areas, it is allowed to drop and create a sense of space and lightness that definitely feels like a stylistic coloration being brought in by the DAC. This surprised me, because I would not have used anything like this description for this combination prior to critically comparing these DACs, but after thinking about it for a few weeks, I believe it is valid in the context of this comparison. It feels like it presents saturation and airiness selectively, but in a generally coherent way to create a very pleasant effect slightly prioritizing space and artfully letting just a tiny bit of detail and tonal richness fall away where it is unlikely to be missed and where its absence will contribute to a sense of drama and atmosphere that is difficult to deny.
The E70 Velvet feels like the creation of a talented student making a very detailed colored pencil reproduction of the painting done by the Qutest. It captures the details well and the colors are mostly natural with only a very slight enhancement of the primary colors reminiscent of the Dayzee, but to a much smaller degree. The difference is in the inability to naturally blend between the colors. In this case, it feels like a pure limitation as opposed to an artful choice to be bold as in the Dayzee. It is a largely accurate reproduction with the same intent as the painting to present reality as it is without editorializing. However, compared to the original, there is a loss of loss of depth and nuance, with an almost totally flattened soundstage and notable over-sharpening of some texture detail. This is not to say that it is bad. It is excellent for what it is, but it is no substitute for the original. It did surprise me that the E70V ended up feeling more like a relative of the Qutest than of the Dayzee, given the shared AKM DNA of the E70V and Dayzee, but the ears hear what they hear!
A few random take-aways
First, I was surprised that the E70 Velvet held up so favorably. Since starting to use higher end DACs exclusively in my head-fi setup, I had really diminished the E70 Velvet in my mind – much more than it deserved. It has a tiny bit of excessive enhancing of detail. This is seldom problematic, but if a recording is very bright, fatiguing, or sibilant, the E70V is not going to do much to fix the situation. It also lacks depth in its staging, but I think this may be a luxury that just demands a high price in the world of headphones, and it is not critically important to the enjoyment of music on headphones. It may also be the case that other DACs will have a better sense of depth at this price, but I would expect that they make sacrifices elsewhere. The E70 Velvet does an excellent job of retaining musicality in the setting of a high-measuring DAC that tries to have a huge set of features in a relatively cheap package. It also perfectly illustrates the diminishing returns inherent in this hobby. It it less than 1/3 the price of the other options, and it will deliver far more than 1/3 of the enjoyment of the others, particularly if one is not swapping back and forth between them.
The second surprise is somewhat related to the first, and it sent me down a bit of a rabbit hole doing research to understand what I was hearing. The E70V punches way above its weight in DSD. It fared very well in the one DSD track in my comparison, and this makes some sense when the topology of the E70V is considered – DSD avoids relying on most of its internal digital processing. Again, I would refer interested readers to the PS Audio blog post on the E70V and DSD that breaks this down beautifully. It should be noted that it only does this when its internal volume control is disabled, so anyone wanting to take advantage of this property of the E70V will need a separate preamp, either standalone or integrated into their amp or done in software prior to hitting the E70V. I would hypothesize that someone using software like HQPlayer to resample to DSD might have a markedly improved overall experience with the E70V, and once I get around to digging into HQPlayer, I plan to test this hypothesis.
A related surprise, but not a positive one, is how poorly the Ares did with handling DSD. This is something I had actually noticed in listening to it previously, but I had hoped the newest firmware that I installed a few weeks prior to this comparison would have fixed this. I don’t think this is a devastating blow to the Ares, as DSD is comparatively rare as a format, and it is easy to resample to PCM, requiring fairly minimal processing power. For people who prefer resampling everything to DSD, I would recommend carefully evaluating this decision when using R-2R DACs. There are, of course, some 1-bit DACs that also use resistor ladders and function by resampling all PCM content to DSD internally, but more traditional R-2R DACs like this one either have a separate circuit for native DSD to analog conversion (like the Ares), bypassing the beautiful resistor ladder that is supposed to be the star of the show, or they simply transcode DSD to PCM internally. This illustrates the importance of understanding each individual device and where it is likely to excel. A format is only as good as a device’s ability to work with it – this is responsible for the sonic differences, not the format itself. Though each format may lend itself to a particular set of advantages and disadvantages, it will ultimately come down to the interaction between the device and the format. Resampling to DSD may be an excellent strategy with something like the E70V, but a terrible idea (at least to my ears) with the Ares.
Another observation I did not expect is that there didn’t seem to be any significant equipment synergy effect in this comparison per se. There was certainly a synergy or lack thereof between content (based either on production values or format) and DAC, but not between the DACs and the two very different amplifiers. The DACs seemed to preserve their relative strengths and weaknesses regardless of amp used. It did, however, seem that the relative differences between them were much more pronounced on the solid state V222 than on the tube-based WA2, where differences just felt subtler generally.
Finally, I know I’m reiterating, but I really was floored by how the Qutest renders extreme detail without compromising musicality or introducing the artifacts I would have reflexively associated with a detailed DAC. I would love to see more innovation moving in this direction as transistors shrink and allow for more to be done in a smaller space with less power or for less money with newer FPGAs. I see potential here for a total disruption of the chip-based DAC market in a way that can stay compact and become more affordable, but of course until there is some competition, Chord has no reason to reduce its prices on these devices. To be clear, I don’t think they’re excessive in this market, but it would be very cool to see a future where developments in FPGA technology would bring extremely refined hifi experiences to smaller and more affordable packages. Oh, and the Qutest has totally broken me from caring at all about the absence of a balanced output on a device, as it turns out that there are other factors that are just vastly more important – someone really should have pointed this out somewhere
Thanks for reading, and I hope that someone finds this helpful! Let me know if you have any questions. Above all, remember to spend time enjoying your music, your system, and the good things you have in your life – at least twice as much as you spend thinking about upgrading!
I originally set out to review the Geshelli Dayzee, and I still intend to, but before I got the chance, I ended up experiencing the Chord Mojo 2. The Mojo 2 really got me thinking about what is possible or desirable in a DAC. My ever-vigilant upgrade gremlin saw the opportunity and made me purchase a Chord Qutest. This left me with 3 DACs in roughly the same price range with very different attributes – the perfect setup for a shootout!
Once I decided to write something up on this for public consumption, it occurred to me that I had seen questions on the topic of the advantages of the Geshelli Dayzee and its dual mono AK4499EX chips over other DACs with the AKM AK4499EX. Unfortunately, I did not have a J2S handy for the comparison, but I do have a Topping E70 Velvet, which I know is different, but it is in a similar price category to the J2S and has a single AK4499EX chip. So, the E70 Velvet became a wildcard 4th entrant in this comparison to give some imperfect insight into the relative standing of a DAC with a single AK4499EX chip versus competition in the next pricing tier up.
Time for some obligatory throat-clearing. In our world, diminishing returns are an accepted reality, and they absolutely do not stop us on our quest for sonic bliss. There is not a bad DAC in this lineup. All comparisons are intended to explore the differences between these products, so individual differences will be pulled out and dissected in far greater detail than might be reasonable or even particularly noticeable in listening to any of these devices casually or even more intently on their own outside of the setting of an A-B-C-D comparison. Also, this comparison is based purely on sound quality, with no consideration of features, appearance, etc. These are all important, and they definitely can affect the experience in different ways, but they will depend very heavily on an individual’s use case. With all of this out of the way –
Let’s meet the contenders!
Denafrips Ares 12th + Iris 12th
For many readers, the Denafrips Ares will need no introduction, but just in case... The Ares 12th (for 12th anniversary) is the successor to the wildly popular Ares II. It is an R-2R resistor ladder-based DAC. The Ares 12th uses a custom USB FIFO interface and FPGA to handle oversampling of incoming PCM to sampling rates in an integer multiple of the source to a max of 1411.2 or 1536 Hz. DSD is not upsampled or converted to PCM but is handled through a separate hardware filtering system specifically included for native DSD support. The Ares 12th also sports a high-quality internal linear power supply and carefully matched resistors to achieve relatively high measured performance, particularly for an R-2R device. The Ares 12th was very quickly superseded by the Ares 12th-I, with the 12th being Denafrips’ primary lower-cost DAC for less than 1 year. Comparisons between the two are available online, but it appears that most of the changes dealt more with quality of life and form factor than sonic performance, so impressions here should be very close to the 12th-I.
The Ares 12th features a nominally non-oversampling (NOS) mode, however, there has been some controversy about whether this feature represents true non-oversampling. This controversy started with the Ares II, and the 12th purported to fix this issue, however, testing from GoldenSound would appear to indicate that, while the results are much more similar to what would be expected by NOS, there are signs that the actions being taken by the DAC do not correspond to what most people regard as NOS. I mention all of this because NOS (particularly the update to “real” NOS with the 12th version) was a major selling point of this DAC and because readers will wonder if it was used in this comparison. I have compared the sound of NOS and OS mode on the DAC with Roon delivering native bitrate files, and to my ears the OS mode is a hands-down winner in sound quality across the board, so the Ares 12th was in OS mode for the entirety of this comparison. The firmware was version 1.8 (the latest available I believe) for this test.
The Ares II received a huge amount of fanfare as a game-changing DAC in the roughly $1,000 price range. When I was looking into picking it up, the Ares 12th was just coming onto the market and appeared to have all of the benefits of the Ares II with a few significant upgrades, so I opted for the new model. It sounded great right out of the box, and it was definitely different from the Matrix Audio X-Sabre Pro MQA that I was using previously. I could hear moments of greatness listening to it, but it definitely didn’t feel as Earth-shatteringly amazing as reviews of the Ares II made it out to be (part of that may have been down to the wonky firmware version that shipped on it initially, and which is no longer even available for download). I was really excited to hear for myself what all the R-2R buzz was about, and I had a distinct feeling I was missing something. That’s when I found the Iris 12th.
The Iris 12th is a DDC, and it appears purpose-built to accompany the Ares 12th in appearance, form factor, and features. I am definitely in the “bits are bits” camp when it comes to digital audio, so I was extremely suspicious of the concept of a DDC, but once again GoldenSound provided the testing that showed that there was a possibility of a measurable difference in actual analog audio output of the chain when adding a DDC. There are multiple possible reasons for this, including more precise timing with the higher quality clock available in the Iris 12th and optical decoupling of electrical signal to prevent noise from bleeding into the DAC from the ground wire of the source being the most compelling to me. One of the other standout features of the Ares 12th over the II was the addition of an I2S HDMI port, which is the preferred connection for the Iris 12th.
I decided to take the gamble on adding the Iris 12th to my system, and it made an enormous difference in refinement and delivering the “magic” I was expecting from the Ares 12th. The combination of the two devices delivered what I was expecting from the Ares 12th based on the hype it was getting. This was reconfirmed when I was moving the Iris 12th and accidentally blew the fuse in it. During the week or so I was waiting for a replacement fuse to arrive, I got to experience its absence after having used it for the better part of a year, and the difference was greater than I remembered (of note, this was still prior to upgrading the firmware on the Ares 12th).
The current version of the Ares retails right around $1,000 and the Iris around $600, though they fluctuate a little because they are sold in Singapore dollars, so exchange rates are a factor. These approximate prices are in USD.
From this point on in this comparison, when I refer to the Ares, I mean the Ares 12th + Iris 12th combination.
Chord Qutest
Many readers will know more about the Qutest than me. I have heard of Chord for many years, but I had never tried any of their products until I stumbled upon a Rob Watts video on YouTube. I had heard him interviewed on a couple of occasions, but in this case, he was giving a presentation at CanJam NYC 2024. (Major shout out to MidFi Guy for putting this on his YouTube channel!) Rob was making a huge amount of sense to me in the way he was describing his process of subjective and objective device testing and refinement. There was something about his presentation that threw me, though. He seemed to have a disdain for the vast majority of the DAC industry, which led me to think that he is either a totally delusional narcissist or perhaps he is onto something. Later it occurred to me that maybe my knee-jerk reaction was a little extreme, and he is an engineer, not a salesman, and he is very passionate about his work... In any case, I was curious about his products, and I started to look for reviews, and virtually all of the people I trust seemed to be over the moon about most if not all of the Chord DACs. I picked up a Mojo 2 and immediately realized I was dealing with something different. This very quickly led me to the Qutest, which felt more “on par” with the level of equipment I have been using.
The Watts philosophy is almost the inverse of what I have learned about what goes into making an excellent DAC. The traditional wisdom, as I understand it, is that converting digital to analog is a relatively straightforward process and has basically been obviated by modern technology, after all, the Apple Dongle has a tiny DAC chip that measures almost perfectly… Oh! Sorry, wrong forum
It is further asserted that all that really matters in terms of sound quality in a DAC is the analog amplification and output stage. Sure, there are different chips (Burr Brown, AKM, Sabre, Cirrus, etc.), but most of them, it is said, sound almost identical, and it is the linear power supplies and analog amplification circuits that make the difference between an excellent DAC and a mediocre one.Chord and Rob Watts, on the other hand, believe that the digital side of the DAC is where the magic happens. The following explanation is ludicrously simple relative to the info available on YouTube, HeadFi, etc., but for the uninitiated, I will attempt to set the stage. Using much higher performing custom-designed FPGA chips enables extreme oversampling of incoming data to more accurately reconstruct the timing of transients. This is combined with careful attention to eliminating or attenuating any sources of noise floor modulation that may vary with the signal. The result is supposed to be extremely precise timing of transients in the music, allowing for a more natural experience of soundstage and timbre. There are also other effects, such as extremely good objective filtering performance with a near vertical drop in signal response around 22 kHz. I hope I have done justice to the basic philosophy of the Chord DACs here, and please feel free to point out anything I missed in the thread below, but if nothing else, this shows roughly what I’m thinking going into this.
One other interesting note with Chord’s devices: they tend to be some of the oldest designs that remain in production at this time. Purchasing the Qutest, a digital electronic product released in 2018, in 2024 feels very odd in a climate where devices are constantly one-upping each other (see above - my Ares 12th was current for less than 1 year). It is not the oldest Chord product still in production. Rob has said (I’m paraphrasing) that this is because competitors are not innovating and are making incremental improvements on a fundamentally flawed design, so there is no pressure to innovate on a faster timeline at various price points. Considering the awards that Chord products rack up year after year and the ability to continually sell units, it seems like he might have a point.
The Qutest is a standalone DAC with dual BNC coaxial (can be used together with certain specialized products to double bandwidth), optical, and USB inputs. The USB input is galvanically isolated to eliminate electrical noise from the source traveling into the DAC. Power is handled by a wall wart delivering 5V DC power via a micro-USB connection. Its only output is via unbalanced RCA connectors. It features 4 filtering settings, Neutral Incisive and Warm – each with the option for adding high frequency roll-off. I did my listening with Neutral Incisive, no roll-off. I believe that this represents the full capability of the device and should allow any special abilities it may possess to shine through. Output voltage is selectable and can be set at 1V, 2V, or 3V at 0 dB. I did my listening at 2V, as this was closest in perceived volume to my other DACs’ RCA outputs. I did not notice a significant difference in sound quality between the settings, just a slight difference in the volume knob position on my amp.
The Qutest currently retails for approximately $1,600.
Geshelli Dayzee
This is Geshelli Labs’ flagship DAC, designed to fit into high-end hifi systems and expand the Geshelli DAC range beyond the extremely popular J2S – which has been heralded as an excellent value for a first serious audiophile DAC as well as a highly aesthetically and sonically customizable device. The Dayzee features 2 of the very highly regarded AKM AK4499EX DAC chips - one for each stereo channel. The only significant audio customization offered is in the form of op amps (a very popular aspect of the J2S, where S stands for socketed, that has carried over to the Dayzee). In the default configuration, there are 8 op amps, 4 per channel, that can be swapped. Unlike with the J2S, these op amps are not obviously associated with specific outputs and cannot be changed individually. All op amps must be swapped if the user desires a different option.
The Dayzee ships by default with OPA1655 op amps. There is an option to swap in the Sparkos SS2590 – the professional bigger brother of the very popular SS3601 discrete op amp. The SS2590 is a larger op amp on a 6-pin professional package with a significantly larger footprint than the DIP8 op amps included by default. In my very early production unit, there was a small hiccup in communication and sockets for the larger pro op amps were not included. This led to an opportunity to break out the soldering iron and attach them to the board myself (more fun than inconvenience in my mind). I only mention this in the service of full transparency, in the extremely unlikely event that this would color the sound (or avoid coloring it) in some way relative to plugging them into a socket. The lovely folks at Geshelli have assured me that all future Dayzee units will have sockets to drop in the pro op amp upgrade regardless of the option selected at purchase.
The Dayzee has dual XLR balanced outputs, as well as dual RCA unbalanced connections. These are all functional simultaneously. Its 6 inputs include USB (galvanically isolated to prevent electrical noise from sources entering the DAC), AES, and 2 each of optical and coaxial (RCA connector). The interface is extremely simple. There is no remote control, filter selection, volume attenuation, etc. All you have is a row of buttons, one of which is a power switch. The others each correspond to a specific input. Notably, DSD is only supported via the USB input, with the others being apparently limited to 192/24 PCM. Power comes from what appears to be a fairly standard wall wart rated for just over 2A of 12V DC.
There are a variety of cosmetic customization options available, and the price of the Dayzee can vary significantly depending on which ones are selected. The base price starts around $1,300 and 8 of the Sparkos SS2590 will run you around $480, though you can get a discount if you order them directly from Sparkos and sign up for their email list. This brings the package up to nearly $1,800 and into a whole new realm of competition for Geshelli.
Topping E70 Velvet
The Topping E70 Velvet (E70V) is the AKM version of the well-regarded E70 that used an ESS Sabre chip. The E70V has one of the AK4499EX chips to handle both left and right channel duties. The E70V, like most of Topping’s offerings, boasts extremely low measures of distortion and noise, delivering measured performance much higher than expected for its price. With the Velvet version of the E70, a level of smoothness (presumably velvety smoothness) was to be added by changing to the AK4499EX without compromising the measured performance. In fact, the measured performance actually improved slightly.
I have been enjoying the E70V for some time now, mostly for near-field speakers hooked up to my computer. I had moved it for a while to my headphone setup, but as I started to get serious about head-fi in the past year, it was relegated back to the computer in favor of the competitors mentioned here. Before this comparison, it had been several months since I listened to it with headphones.
For inputs, the E70V sports coaxial, optical, USB, and Bluetooth. Output is available fully balanced via XLR out or unbalanced over RCA. It also has a wide variety of features, including selectable filters, remote control, digital domain volume attenuation allowing for function as a preamp, adjustable voltage output on XLR (4V or 5V) and RCA (2V or 2.5V). The 4V and 2V output levels were used in this comparison. Power is handled by an internal SMPS, connected by a standard 3 prong cord.
One interesting aspect of the AK4499EX chip is that it really is a set of two chips, one analog and one digital. This allows for better isolation of the analog stage from the digital conversion, an inherently noisy process. I don’t want this to be confused with the Dayzee, which has two of the AK4499EX, with each of these consisting of two chips (analog and digital). This dual-chip configuration is part of what raises the price on devices featuring the AK4499EX over their counterparts. There are significant differences between the E70V and the Geshelli J2S (also available with a single AK4499EX in its top configuration, which is notably pricier), but this is the closest I have available for comparison.
One final fact about the E70V will be especially exciting to the value-oriented DSD enthusiast. The E70V, when run in pure DAC mode (with volume attenuation disabled), is able to bypass the delta-sigma conversion process completely, allowing native handling of DSD. When volume control is enabled, the DSD signal is converted to PCM and fed through the same D->A conversion pathway as PCM sources. The native handling of DSD is purported to give the AK4499EX an advantage in DSD playback, but only when correctly implemented. Some devices with the AK4499EX do not take advantage of the ability to bypass the PCM conversion step. The E70V was running in pure DAC mode for this comparison to take advantage of this feature. Tom Gibbs has a fantastic piece detailing this aspect of the E70V on the PS Audio website - https://www.psaudio.com/blogs/coppe...tive-dsd-boom-with-the-topping-e70-velvet-dac
I believe that the E70V is one of the least expensive implementations of the AKM flagship AK4499EX at $450.
The Setup
I am not a professional reviewer, and this test was not intended to be scientific, but I did try to control as many of the variables as feasible. I do believe that the power situation in my home is a bit prone to noise and voltage fluctuations, if the frequent, sudden changes in brightness of the lights are any indication. To give each device its best chance to shine, I used a Pine Tree Audio combined AC / DC power filter in front of a Furman AC filtering power strip. Two of the DACs (Ares and E70V) have fully internal power supplies, so this is all the help they’re getting. The two with wall warts got a little extra attention in the form of iFi iPower X adapters in their respective voltages in place of their included wall warts.
I listened to 6 tracks with which I’m quite familiar and took notes. I listened to a track on all 4 DACs all the way through prior to moving to the next track. I kept the order of the DAC rotation the same, but I started with the next DAC in the rotation on each track, so each one got the benefit of being the first to play at least one track.
All listening was done on my ZMF Atrium with universe suede pads and stock mesh connected by ZMF 2K Copper cables (1/4 inch for the single ended amp and 4 pin XLR for the balanced one).
The first 3 tracks were tested on the Woo Audio WA2 amplifier – an output transformer-less, single-ended triode tube amp. Tubes used were NOS Mullard-made E188CC drivers, NOS Chatham 5998 power tubes, and NOS Mullard-made EZ80 rectifiers. All of these tubes are in good condition and have between 100 and 300 hours play time depending on the pair. This amp has also been modified with the interstage coupling caps having been replaced with Mundorf Supreme EVO Silvergold (no oil). Each DAC was connected to the WA2 by a set of Kimber Kables Hero RCA cables.
The last 3 tracks were tested on the Violectric V222 amplifier – a fully-balanced solid-state amp. All 3 user-accessible op amps have been replaced with Sparkos SS3602 discrete op amps. As this amp uses a fully balanced topology, and as I understand it, this is the preferred way to connect devices to it, I used World’s Best Cables Mogami star quad XLR cables to connect the 3 balanced DACs, and I used the Kimber Kables RCA from the other test to hook up the single-ended-only Qutest.
Volume was not scientifically matched. I just sought to start slightly low and increase to the loudest comfortable listening level as quickly as possible with each track (I believe I tend to listen somewhere on the threshold of moderate to loud, but I don’t have the setup to measure SPL to be sure).
4 tracks were streamed from Qobuz and 2 from my files. Rips of the 2 tracks from my own physical media collection were done with great care to get bit perfect copies – one PCM from Red Book CD and one DSD from SACD.
Playback was handled by Roon. All tracks were played in their native format with no DSP (fixed volume in Roon, purple “lossless” label present). My Roon Rock NUC streamed all tracks over WiFi to a Raspberry Pi 4 running RoPieee. The Pi was connected to all 4 DACs simultaneously by USB using generic cords provided with the devices or similar equivalent. Note that all of the options have some sort of galvanic isolation except the E70V, so I used a Topping HS02 between the E70V and the Pi to galvanically isolate, so noise from this source should not be problematic. I assume most people looking at DACs in this level of depth will have an understanding of the importance of isolating from an electrically noisy source, so I didn’t want to handicap the E70V, as most people on HeadFi will probably use a streamer built for low noise or an isolating strategy if using a DAC that does not have galvanic isolation.
The Comparison - Listening Notes (presented in order of listening)
Round 1 – Woo Audio WA2 SET OTL Tube Amp
Saturnine – GoGo Penguin – Qobuz 48 kHz 24-bit
Dayzee – Synth at the beginning of the track is clean and well-defined, and the low end is particularly powerful. Moving to string bass, higher notes start to have a slightly harsh character – detailed, but slightly grainy or over-sharpened. Timbre is good, nowhere is it particularly problematic in this track. Piano has excellent note weight – not overbearing or excessively emphasized, but very enjoyably engaging. Left-right sound stage is presented well, but not much depth is noted – everything feels intimate. Tonal balance is good, but it feels like there is a very slight bass boost.
Ares – The reverb and ambiance created by the opening synth are immediately arresting. Bass is adequate, not exaggerated. The tonal quality (as opposed to transient) of the snare drum is very noticeable. Timbre of the piano suffers relative to Dayzee – leading edge feels smeared, and it sounds like it could be a piano or synthesizer. There is a real sense of space right from the beginning and instruments are spread evenly in the soundstage with no “blobs” emerging. There is a modest sense of depth – particularly noted on snare drum. There is a slight feeling of an element of height elevation with piano. There are no harsh or grainy tones. Overall, a pleasing balance of note weight and air.
E70V – Movement of the synth in the beginning of the track around the soundstage is instantly apparent. This is not something I had appreciated with the first 2 DACs. The standup bass has a very noticeable sense of texture, but it feels slightly exaggerated – not fatiguing, but doesn’t seem as natural as the others. Good sense of timing and rhythm – feels faster than the first 2. Feels more detailed overall than the first 2, unclear if exaggerated overall or appropriate to the recorded material. No significant depth is appreciated at all. Tonal elements of percussion are almost completely absent here.
Qutest – Air and sense of ambiance are very noticeable right away, similar to Ares, maybe slightly more. Movement of synth is appreciated again here, but not as dramatically as with the E70V. Some texture is present on upright bass – not as much as with Dayzee or E70V – no harshness nor exaggeration at all. Timbre of the piano and percussion are absolutely convincing. Note weight is adequate, but feels lighter than the others, unclear which might be closer to the intent of the recorded material. Detail is highest of all DACs tested here – no exaggeration, harshness, nor etching. Left-right spacing is good. Depth is still modest on this track, but more than with any of the other DACs
Paper Tiger – Beck – MoFi CD 44.1 kHz 16-bit
Ares – Intro vocals feel very intimate. There is a great deal of air and separation around all instruments. There are subtle details present in the electric guitar. Handles busy passages well throughout the frequency spectrum – maintaining good sense of air and detail when things are complex. Reverb is pronounced. There are harsh frequencies in this recording to my ears, and they are tamed adequately by the Ares. Generally pleasing and laid back but leaves something to be desired in terms of slam and macro-dynamics.
E70V – Reverb feels very sharply detailed – sound does not feel like it diffuses through space. There is significant detail in this presentation, sounding slightly dry. “S” sounds in vocals are dangerously close to excessively sibilant, but not quite crossing the line. Soundstage is wide, but there is absolutely no depth – it has the sense of being on a cellophane film cutting through my head from ear to ear – not that it is harsh, but that there is no sense of depth whatsoever beyond the line between my ears. There is a buzzing character to the timbre on string instruments that I would attribute to overly etched detail. Still not particularly fatiguing. Presents a generally enjoyable experience in spite of some of the shortcomings that make it slightly less natural.
Qutest – Reverb on the opening vocals is subtle, detailed, sounds like a natural reaction to a space in which they are being sung. The bass guitar riff is a great deal easier to follow with all articulation present – this is a significant departure from prior 2 DACs. Sibilance in the recording is clearly noted but feels as though it has somehow been rendered inoffensive. Sounds of guitar strings responding to the individual frets as the musician’s fingers slide across are present in a way I don’t recall having heard before in this track. Complex passages feel effortless. Timbre of cymbals is particularly convincing. Micro-dynamics are excellent. Macro-dynamics are fine, but not particularly stunning.
Dayzee – Less sense of space and reverb on intro than the other DACs. There is a pleasing enhancement of note weight present in vocals that was not there for the other DACs. Detail is great, not forced. Sense of depth is minimal, but more than E70V. Imaging and left-right separation of instruments, as well as ability to pick them out of the mix are very good. Timbre of the percussion is notably good, but not quite on the level of the Qutest. Feels as though there is a theme throughout this track of note weight / tonality being enhanced over timbre and general technicality.
Creepin’ In – Norah Jones w/ Dolly Parton – Analogue Productions DSD (SACD)
E70V – Nice sense of rhythm and timing throughout – toe-tapping goodness. Detail is good overall. Maybe a tiny bit of excessive grain applied to Norah’s voice. Dolly’s voice sounds perfectly natural here. The bass line can be followed, and intensity is appropriate, but it is not particularly well-articulated, and there is minimal timbre information in the lower instruments. Melody creates an emotional response that is noteworthy, particularly given what I would have expected from this DAC. There is a sense of spatial depth that emerges in this track, particularly in the vocal harmonies.
Qutest – Norah’s vocals are presented with a striking level of detail but with no sense of artificiality or harshness at all. Similarly, Dolly’s voice is more textured and never harsh. Both sets of vocals feel like masterful blends of tonality and texture, presented naturally. The bass guitar is noticeably better articulated than with E70V. Generally, the instruments, like the vocals, have a tremendous balance of tonal information and realistic timbre. It is genuinely difficult to take notes here, as this is such an enjoyable listening experience.
Dayzee – There is a bit less air / separation of instruments relative to the 2 previous DACs in this track. Similarly to the 2 prior tracks, the Dayzee delivers a slightly elevated bass register. Detail is lost in the lower instruments here, similar to E70V. Note weight is enhanced relative to the other 2 DACs. Any harshness detected in Norah’s voice is totally smoothed out and pleasant here. Dolly, on the other hand, sounds a little bit distant and muffled. Timbre of percussion is good but fades in busier passages. All of the vocals have enhanced tonality – bringing some of the background vocals forward. There is some perceptible depth, but it is modest.
Ares – Even with raising the volume, the entire presentation sounds recessed, as though there is a sense of depth being applied, but to every element simultaneously. Backing vocals sound as though they have been pulled forward in space relative to the other DACs. There is a sense of things moving around in space as the track gets busier in a very odd and disorienting way that is difficult to describe. There is some very interesting detail in the tonality of the drums that I did not notice on other listens. Generally, the spatial presentation is quite distracting and unnatural. The tonality is pleasant, but I cannot get into the song at all. I did not do a second critical listen, but I tested a minute of the track using resampling in Roon to 192/24 PCM and the performance was much better – in line with the quality level that would typically be expected for this DAC.
Round 2 – Violectric V222 Fully Balanced Solid-State Amplifier
The Hardest Cut – Spoon – Qobuz 96 kHz 24-bit
Qutest – The clapping starting early in the track and continuing behind the drums is something I had not differentiated as a specific element of the percussion until listening on this setup. Very fine details of 3 distinct guitars are clearly noticeable and can be easily identified even as all are styling simultaneously. A richness of vocal harmony in the pre-chorus is particularly stunning. Staging is intimate overall, but during the bridge and solo, space expands out before snapping back in in a way that feels intentional in the production of the track. Excellent balance of musicality and technicality.
Dayzee – Clapping at the beginning of the track is difficult to differentiate as the track builds intensity. Good detail is present in the guitars, and the note weight and tonality are more impactful relative to Qutest. Vocal harmonies in the pre-chorus collapse into a single tone, with a loss of distinct detail between the voices. This track is generally intimate, and this is especially apparent here. No depth is noted at all. There is no harshness, etching, or excessive detail. Though not particularly technical, this is a very musical, enjoyable listen.
Ares – Clapping at the beginning of the track stays clear as other percussion picks up. There is a dramatic sense of air and reverb on the clapping, creating a more spacious feel than the 2 preceding DACs. Depth and layering are also very evident in the vocals. The guitars are all detailed and can be differentiated, only slightly less technical than the Qutest. There is a particular sense of groove to the bass line on this listen that is well-articulated and fun to follow along with. The tonal richness is not quite as high as the Dayzee, but there is notably more detail. Qutest is more detailed than Ares and has better timbre, but Ares has a larger overall sense of space on this track.
E70V – Immediately noticeable increased intensity in the bass. Though elevated relative to the others, bass is driving and pleasant. Clapping is clear and mostly preserved. Interestingly, the reverb effect on the clapping comes through clearly, and yet there is absolutely no sense of perceptible space or depth generated by it – an odd effect. The level of fine detail on the guitars is surprisingly high and goes toe to toe with the Qutest. The overall presentation is tonally rich and musical, but spatially flat.
The Unclouded Day – Randy Travis – Qobuz 44.1 kHz 16-bit
Dayzee – Immediately musical and tonally rich. This enhancement of note weight is welcome in this somewhat bright and very busy production. Timbre is very good. There is a bit of perceivable depth, particularly in the layering of the vocals. There is good air and space around the many instruments playing, and it feels like they can be explored without any of them being pushed forward excessively. There is a slight enhancement to the lower registers that is noticeable and has a balancing effect on this track. Great listen overall, technically and tonally.
Ares – Relative to the Dayzee, it is immediately apparent that the timbre is more natural on the instruments. The bass drum feels somewhat less impactful. The upright bass still has a sense of driving rhythm and toe-tap-ability. Randy’s voice relative to the Dayzee feels lighter and airier – slightly ethereal. Individual voices in the vocal harmonies are a bit difficult to pick out, but they feel cohesive and musical in the way they combine.
E70V – Compared to the first 2 DACs, there is a sense of dryness, but it is subtle, and I don’t know if it would be noticeable if not for directly following the other 2. Generally, still musical and engaging tonally. Soundstage: flat as a pancake. The presentation is intimate, which is surprising after the other DACs. The vocal harmonies collapse together, and voices cannot be differentiated. The instruments feel detailed, but they do not have space nor air. It seems like etching and excessive sharpening to create the perception of detail destroys some real detail here. I suspect this is because of the brightness of the mix – I’m not sure what year it was recorded, but it strikes me that the producer had the 80’s sensibility that he had never heard a mix with too much treble. This is not good for synergy with the E70V. It is not terrible, but this is a surprisingly striking weak point for the E70V.
Qutest – Timbre and instrument spacing in this track the best of the 4 by far. There is plenty of room throughout the entire track to explore the details of each instrument or mentally zoom out and appreciate the natural and cohesive overall production. The high treble is still present, but similar to the sibilant S sounds on the Beck track, the Qutest does its trick here where it honestly presents the information but somehow keeps it from being offensive. The note weight in Randy’s voice is lowest with Qutest of the 4, but the texture and timbre are totally natural and effortless. Though this is my 4th listen in a row, it is clearly the most emotionally arresting (this is a song that has affected me emotionally since I was a child). I actually feel a bit choked up while listening, which I was not expecting.
Chocolate Chip Trip – Tool – Qobuz 96 kHz 24-bit
Ares – Reverb is in full effect throughout the track. Though this track is almost exclusively percussion, there is a significant and noticeable amount of tonal richness. Lots of air present around the various elements. The left-right soundstage is smoothly rendered and gapless. Overall presentation is coherent and timbre, particularly on the acoustic drums, is very good. Bass drum has good slam without feeling excessively enhanced or washed out. There is noticeable depth in the center of the soundstage where multiple elements overlap.
E70V – Tone is not as pronounced with the E70V, and there is less air around instruments. There is an enhancement of texture, and the buzzer-like elements of the electronic instruments are brought to foreground. Depth is modest, but this presentation is not totally flat like some others with this DAC. Timbre is decent on acoustic drums, but not as convincing as Ares. Left-right tracking is smooth with no blobs, but the overall soundstage is more intimate.
Qutest – For the first time, after having used this as a test track since shortly after it was released, I notice the slight out of phase vibrato (if that’s the right term) between two of the tones in the beginning – similar to what you get when you are tuning a guitar and are almost there, but not quite. The timbre and tonality are balanced and very natural. There is a very large amount of fine texture, but it does not feel etched, enhanced, or harsh. The lateral localization on this listen is the best I have ever heard from any setup on this track (not that I have ever had any that are much higher end than these) and the ability to pick the elements apart is on a level I have not experienced. Technicality is through the roof. The low end of the frequency spectrum feels thin relative to the prior 2 DACs - though I wonder if this is just absence of artificial enhancement, given the exceptional technical performance on this track. There is still significant slam on deliberate hard drum strikes. It is difficult to overstate how blown away I am by the detail and technical performance on this track. This track can make almost any DAC sound pretty awesome, but the Qutest really feels like it was taken to another level on this one.
Dayzee – Note weight and tone are notably enhanced here relative to all of the others. There is less air around the instruments, but there is some noteworthy depth present in the center of the soundstage and great, even left-right separation. Timbre is less convincing than all of the other DACs. There is a washing-out effect on tone in the acoustic drums from the low-mid register and down. Though it is not a particularly technically impressive performance relative to the other DACs here, it is extremely fun. The slam is off the charts, and it is not hard to imagine why someone might prefer the enhanced macro-dynamics and associated giggle factor over technicality or even optimal tonality. There is still a high level of technical proficiency here, but the slight bass boost buys you a surprising amount of ferocity at the cost of a bit of precision.
Conclusions aka TL;DR
This project has taken longer than I expected, particularly finding time for the writing, so I have had a lot of time to think about it. I have decided to break this conclusion into 2 parts: first, the general characterization of the DACs in relation to one another; second, a collection of individual impressions and takeaways that I think might be valuable or that surprised me in this process. This is also a good place to reiterate that all of these DACs are competent in all settings, and strengths and weaknesses are being highlighted to differentiate between 4 very reasonable options.
The Synthesis
At the very real risk of being painfully cliché, I feel like the most coherent analogy I can craft to compare these 4 DACs is a visual one – Imagining the performances of each of these DACs as artists’ renderings and the recordings themselves being the actual scenes being represented. I think the general differences can be likened to a choice of artistic medium and style.
The Qutest is an extremely realistic oil painting. The presentation is impeccably detailed, life-like, and without excessive interpretation. In general, it seeks to capture the essence of the scene, almost like a photo. The technical presentation is absolutely stunning, and prior to experiencing the Qutest (and the Mojo 2), I would have thought that this description would be incompatible with musical enjoyment – at least for me. I have been a longtime lover of vinyl, and for the past 5 years I have also been enthralled with tubes, so I have a reflex to equate high levels of detail with a lack of soul and a bright, edgy, harsh experience – like I had with the Matrix X-Sabre Pro. The Qutest has completely disabused me of this notion, and I am simply in love with the sound of it. I cannot believe that there has not been more of an effort by competitors to follow Watts’ approach and eschew the received wisdom of enhancing the output stage while relying on an off-the-shelf chip. To the extent this is happening, it appears to be in R-2R DACs, but Chord is proving that abandoning microchips is not the only way to get to a different sound. I think anyone with a level of audiophile sensibilities, on hearing the Qutest or Mojo 2, would have to acknowledge that there is something different going on. I believe it is a line of inquiry that deserves to be fully explored.
The Dayzee comes across a bit more stylized. To me, it feels like it is rendered in bold, vivid, primary color acrylics. There is a sacrificing of fine detail in the service of tonal richness and impact. I think of opaque, striking reds, greens, blues, and yellows with thick black and white lines separating them, or perhaps the colors abutting one another, but with minimal if any blending. This DAC is competent in timbre and spacing, but unremarkable at its price point relative to the competitors here. Depth and texture may be diminished, but there is a striking sense of tonality that can enhance the emotion being communicated in a melody regardless of valence. By avoiding adding any over-etched enhancement of detail or harshness, listeners can be content to relax and let the Dayzee put the melodies and harmonies front and center for accessible enjoyment. The most fun features of songs with intense bass and slam will be even more thrilling here. There is enough detail and technicality that it will not feel like anything is being missed unless it is specifically compared to one of a handful of more technical DACs.
The Ares 12th + Iris 12th combination is a realistic rendering done with pastels. There is a sense of realism and technical detail, but there is also an enhancement of air and a generally ethereal sense being imparted to the music. Tonal density is variable – most of the time it is opaque and deep, but in certain areas, it is allowed to drop and create a sense of space and lightness that definitely feels like a stylistic coloration being brought in by the DAC. This surprised me, because I would not have used anything like this description for this combination prior to critically comparing these DACs, but after thinking about it for a few weeks, I believe it is valid in the context of this comparison. It feels like it presents saturation and airiness selectively, but in a generally coherent way to create a very pleasant effect slightly prioritizing space and artfully letting just a tiny bit of detail and tonal richness fall away where it is unlikely to be missed and where its absence will contribute to a sense of drama and atmosphere that is difficult to deny.
The E70 Velvet feels like the creation of a talented student making a very detailed colored pencil reproduction of the painting done by the Qutest. It captures the details well and the colors are mostly natural with only a very slight enhancement of the primary colors reminiscent of the Dayzee, but to a much smaller degree. The difference is in the inability to naturally blend between the colors. In this case, it feels like a pure limitation as opposed to an artful choice to be bold as in the Dayzee. It is a largely accurate reproduction with the same intent as the painting to present reality as it is without editorializing. However, compared to the original, there is a loss of loss of depth and nuance, with an almost totally flattened soundstage and notable over-sharpening of some texture detail. This is not to say that it is bad. It is excellent for what it is, but it is no substitute for the original. It did surprise me that the E70V ended up feeling more like a relative of the Qutest than of the Dayzee, given the shared AKM DNA of the E70V and Dayzee, but the ears hear what they hear!
A few random take-aways
First, I was surprised that the E70 Velvet held up so favorably. Since starting to use higher end DACs exclusively in my head-fi setup, I had really diminished the E70 Velvet in my mind – much more than it deserved. It has a tiny bit of excessive enhancing of detail. This is seldom problematic, but if a recording is very bright, fatiguing, or sibilant, the E70V is not going to do much to fix the situation. It also lacks depth in its staging, but I think this may be a luxury that just demands a high price in the world of headphones, and it is not critically important to the enjoyment of music on headphones. It may also be the case that other DACs will have a better sense of depth at this price, but I would expect that they make sacrifices elsewhere. The E70 Velvet does an excellent job of retaining musicality in the setting of a high-measuring DAC that tries to have a huge set of features in a relatively cheap package. It also perfectly illustrates the diminishing returns inherent in this hobby. It it less than 1/3 the price of the other options, and it will deliver far more than 1/3 of the enjoyment of the others, particularly if one is not swapping back and forth between them.
The second surprise is somewhat related to the first, and it sent me down a bit of a rabbit hole doing research to understand what I was hearing. The E70V punches way above its weight in DSD. It fared very well in the one DSD track in my comparison, and this makes some sense when the topology of the E70V is considered – DSD avoids relying on most of its internal digital processing. Again, I would refer interested readers to the PS Audio blog post on the E70V and DSD that breaks this down beautifully. It should be noted that it only does this when its internal volume control is disabled, so anyone wanting to take advantage of this property of the E70V will need a separate preamp, either standalone or integrated into their amp or done in software prior to hitting the E70V. I would hypothesize that someone using software like HQPlayer to resample to DSD might have a markedly improved overall experience with the E70V, and once I get around to digging into HQPlayer, I plan to test this hypothesis.
A related surprise, but not a positive one, is how poorly the Ares did with handling DSD. This is something I had actually noticed in listening to it previously, but I had hoped the newest firmware that I installed a few weeks prior to this comparison would have fixed this. I don’t think this is a devastating blow to the Ares, as DSD is comparatively rare as a format, and it is easy to resample to PCM, requiring fairly minimal processing power. For people who prefer resampling everything to DSD, I would recommend carefully evaluating this decision when using R-2R DACs. There are, of course, some 1-bit DACs that also use resistor ladders and function by resampling all PCM content to DSD internally, but more traditional R-2R DACs like this one either have a separate circuit for native DSD to analog conversion (like the Ares), bypassing the beautiful resistor ladder that is supposed to be the star of the show, or they simply transcode DSD to PCM internally. This illustrates the importance of understanding each individual device and where it is likely to excel. A format is only as good as a device’s ability to work with it – this is responsible for the sonic differences, not the format itself. Though each format may lend itself to a particular set of advantages and disadvantages, it will ultimately come down to the interaction between the device and the format. Resampling to DSD may be an excellent strategy with something like the E70V, but a terrible idea (at least to my ears) with the Ares.
Another observation I did not expect is that there didn’t seem to be any significant equipment synergy effect in this comparison per se. There was certainly a synergy or lack thereof between content (based either on production values or format) and DAC, but not between the DACs and the two very different amplifiers. The DACs seemed to preserve their relative strengths and weaknesses regardless of amp used. It did, however, seem that the relative differences between them were much more pronounced on the solid state V222 than on the tube-based WA2, where differences just felt subtler generally.
Finally, I know I’m reiterating, but I really was floored by how the Qutest renders extreme detail without compromising musicality or introducing the artifacts I would have reflexively associated with a detailed DAC. I would love to see more innovation moving in this direction as transistors shrink and allow for more to be done in a smaller space with less power or for less money with newer FPGAs. I see potential here for a total disruption of the chip-based DAC market in a way that can stay compact and become more affordable, but of course until there is some competition, Chord has no reason to reduce its prices on these devices. To be clear, I don’t think they’re excessive in this market, but it would be very cool to see a future where developments in FPGA technology would bring extremely refined hifi experiences to smaller and more affordable packages. Oh, and the Qutest has totally broken me from caring at all about the absence of a balanced output on a device, as it turns out that there are other factors that are just vastly more important – someone really should have pointed this out somewhere
Thanks for reading, and I hope that someone finds this helpful! Let me know if you have any questions. Above all, remember to spend time enjoying your music, your system, and the good things you have in your life – at least twice as much as you spend thinking about upgrading!
. 
I borrowed a chord tt2 + chord mscaler for a few days…. Which lead me in buying it
. But I do say, only buy if you can, not if it is your last cash, it is still a luxery hobby
.