Gustard DAC-X26 impressions and discussions

I’ve played with this dac roughly for one week. I’d share my impressions by now, before I will start my evaluation of Matrix Sabre-pro tomorrow.

Gustard has claimed that he came up with the own implementation of digital filter to improve DS quality. I mostly like such approaches. Yes “mostly” -- when they worked, it could raise the total quality into a whole another level like DM nailed with Convert-2. However, wrong engineering can also give birth to a complete garbage like (recent) AGD shits. Their r2r dacs really made me more cautious toward any original implementations. Btw agd 9038 also sucked. Starting to suspect Kingwa’s incapability to handle digital circuits.

For this reason, I didn’t hope much for Gustard at first, as there were already tons of BS with his products. The most notorious example was DAC-X20 (9028-based) combined with weird mods. I could get my ears on modded x20 at several different circumstances. To my ears, it wasn’t as f'ing as agd. But lacking coherence and uncontrolled “sabre” were enough for me to walk away. I still believe much cheaper Modi 3 and DX3 pro destroy X20 no matter whether or not modded.

Then why did I decide to spend time with Gustard again? That’s because of my audiofool buddy (we call ourselves like this) who owns x26. One day he came to my place to hear convert-2. We enjoyed dac talk as well. At the end of intensive but directionless chats, he said “Hey bro, you should hear my gustard. You must like it.” If that’s from a random person, I’d just ignore it. But as from one who knows my tastes, the statement a little motivated me. Plus, he kindly loaned me his dac.

And since then, I’ve been seriously testing x26.

Shortly speaking, it’s legit.

Let me be clear. It’s NOT a giant killer. Regard any dude arguing x26 beats yggdrasil as a deaf, liar, or shill. But this is the first solid performer that I think can replace convert-2 without too much loss. Final decision will be made after I test Matrix X-Sabre pro loaner unit (I’ll get it likely tomorrow) though.

Before X26, the most impressive 9038 dac was smsl d1. In short, it showed me well-controlled sharpness. Thus, as a 9038 dac, I was originally expecting x26 could sound like soekris 1421, but it didn’t. Rather, it mimicked Gungnir Multibit(a1) pretty much.

I will explain more details later but here are some of the highlights:

• USB was unusually bad. SPDIF coax input was ok (nothing wrong). And I like toslink optical the best with this dac, which is also unusual.
• There are several options of jitter controls (via PLL filter). “Cdt opt” worked better than any other filters. No idea what that actually means. Anyway compared to bypassing, the change was simply dramatic. I’d say the change is as drastic as Convert-2’s master clock mode (in comparison with other clock modes).
• Great transience. Nearly on par with convert-2. Maybe one or a couple of ticks overshooting. I’d say 95% accuracy of convert-2.
• Low level plankton are decently resolved, with a hint of over-exaggeration. Less (truly) resolving than yggdrasil 2 and less vivid than convert-2.
• Pretty neutral tone. Maybe only slightly tipped up on top -- nevertheless it doesn’t touch my nerve with marginal or sibilant recordings.
• Bass is less meaty than Gumby1 (from the memory) but more so than convert-2. Less slamming than convert-2. But still very enjoyable.
• Staging is clearly better than Convert-2 (when tested with both JLH+Lore speakers and BW+Ether 2). Nice layering and distance somewhere between Yggdrasil2 and Gumby1. But not as fully 3d-ish as Schiit multibits -- objects placed at noticeably different coordinates on z-axis but each object is very thin like a sheet of paper.
• Less veil than convert-2 at a cost of less blacker background. If convert-2 is like iPhone X display, x26 is iPhone 8 screen. And with this analogy, yggy2 is like an eizo lcd monitor.
• Speaking of aggressiveness, X26 is more forgiving than the average 9038pro dacs I’ve tested so far. D1 was almost the median among the bunch, and similar with 1421. X26 is more in line with Gumby1. Maybe a tad brighter. Clearly less incisive than convert-2.
• Overall X26 sounds like Gumby1 souped with a good traits of Convert-2 and the good Sabre (9018 and 9028 automatically disqualified). Downside is still a bit artificial sounding instead of true naturalness of Yggdrasil A2.

 

Great review. I was wary of Gustard but keep reading very good reviews of the X26.

What filter was yours set to? VIVID (default) / COMPOSITE / GENTLE? I suppose default?

 

Gentle was the least preferred. Preference bw vivid and composite is much less clearer. To my rig and computer-sourced signal, vivid seems the best so far.

Note that any of filter names do NOT reflect how they actually sound. lol. Composite sound more vivid. Gentle sounds a bit blurred and unfocused. Vivid adds somewhat buttery nuance, which I like but some may hate.

 

That is just want I wanted to know

 

very interested to hear more impression from this dac!

 

X26 uses an ESS usb chip (rather than XMOS thing), which doesn't seem to work for me. Not very sure if wyrd can improve this situation. I got a feeling that it's not a matter of noise -- but something weird is going on inside the chip.

I'm using matrix xspdif now (recently changed from eitr). Any decent converter to either spdif or toslink will work because X26 can reclock digital signal fairly well. Can recommend xspdif, su-1, or eitr for such purpose. Just avoid usb input if possible.

 

Not very important info. But I've just weighed this dac. I initially felt x26 quite heavy. The scale confirmed it.

DAC-X26 weighs 13.0 lbs, which is noticeably heavier than Convert-2 (10.9 lbs) and Gungnir (11 lbs). Note that the two dacs of comparison have larger form factor. X26's product dimension is 13.0 in(W) x 2.6 in(H) x 10.3 in(D). Most likely the weight is due to dual power transformers and thick-ish aluminum plates.

Looking at the bright side, x26 is still transportation-friendlier than Yggdrasil (25 lbs)..

I am feeling that my Matrix XSP loaner evaluation will be a thorough shootout between two capable Chinese dual-9038pro dacs. Need to return x26 next Monday though.

 

After serious 2-day long shootout during this weekend, I finally decided to go for X26 (indeed order placed already).

Some of comparative impressions can be found here (in X-Sabre Pro thread): https://www.superbestaudiofriends.o...ream-of-consciousness.7598/page-2#post-252006

XSP surely has more potential and does a lot of things in the very right ways: tonality, staging, honesty, etc. But whenever switching from xsp to x26, I became immediately happier for both testing tracks and general listening. More incisive and slamming (still less so than Convert-2), tastefully exaggerated small dynamics, and lovable transience and attacks. More tactile texture. More-resolving-ish. Blacker background and less veiled sonic images are noticeable too. I’d repeat some of those might be done in wronger (exaggeration and falsification) ways. Kinda guilty pleasure there.

I am inferring that the difference largely comes from my sub-optimal digital transport, because I could feel most of such degradation when x26's jitter control bypassed (to different extent though). Likely xsp is very delicate and can scale with higher quality transports. By comparison, x26 seems to “equate” (as opposed to scale) whatever I throw in. Capitalism vs commercialism in the dac world!

I will be able to recommend XSP to any enthusiasts any day, particularly if they’re coming from well-equipped hardcore hobbyists. But for me, for now, and with my unwillingness to upgrade transports by foreseeable future, can't deny desire to get X26 for myself -- it just works for me as much as I want.

PS. Last night, I thought finally found the best setting for XSP that fantastically worked for me. Got truly excited… until found “the other dac” was connected to the amp. Haha. Probably that moment, decision was made.

 

Nice write up @Vtory. I've been reading a lot of good things about ES9038 DACs.
I wonder how the NFB-7.77 stacks up against it, being another dual mono ES9038 implementation. Can't find any reviews for it.

 

@gridmaster

I have been consistently disappointed by Kingwa's dac products. The sole exception was old master 7 (pcm1704-based) dac years ago -- not competitive for today's standards but listenable as I recall correctly. It seems that he kinda prefers less resolving and syrupy sounds. Master 7 was a borderline of good balance between such tone and technicality. Unfortunately his recent r2r products are .. in short I will avoid them at any moment even if they are given for free. Great example of bad warm poos.

Surprisingly Agd's ess dacs also seem to share this character in bad ways, regardless of very different dac chips. I haven't heard the one you mentioned but older 9038 dac completely demotivated me. It sounded rather similar with agd r2r. My bet is that his analog circuit is inherently problematic -- or it just didn't work for my high-end standards. Anyway modi 3 was more resolving, more slamming, and much righter by comparison.

 

@Vtory Interesting, thanks for the input. I personally own a 1704 m7 and yggyv2.
I've read on forums that the latest r2r customs he has are good, but I wouldn't know never heard one.

I know many people reviewed and listened to the 1704 m7 before Kingwa came out with a DSP firmware revision for it.
Completely changed the sound of the dac, zero syrup anymore (and I agree with its original firmware was very much like that).

Unfortunately another thing many people overlook when reviewing the unit, is the ability to disable the onboard reclocker.
My source is a Lynx aes16e, and the difference is night and day having PLL turned on/off.

It still suffers from bloated bass, but not to any extent that bothers me. It's actually more punchy and macro dynamic than my Yggdrasil. I'd say it's a more "fun" listen. While the yggyv2 retains better tonal balanced, micro dynamics, soundstage, and a hair more resolution. M7 actually has a blacker background tho.

Not to derail the thread, I'll leave it at that since it's a dac-x26 thread.

 

If you own M7 with latest firmware and a good transport (I've not used lynx aes but heard a lot of good things from trusted ears), just ignore my m7 assessments -- mine are from usb input. Even under such condition, it was the only agd product I liked. Very lovely midrange tone for most vocals. Seriously considered to get one when attractive prices popped up early last year (sadly couldn't make it).

 

This is both sad and enlightening when it happens. In this case, it worked in your favor!

 

I have an Audio-GD NFB7.77 and a Gustard X26. I have not done a proper comparison of the two yet but I will try to share some early thoughts here. I will follow up at a later date with an organized A-B comparison. I can comment mostly on usability features and a little bit on sonic impressions for now.

Spoiler: Not entirely necessary, but some background

I've been running the Gustard at home on my stereo system and the NFB7.77 on my headphone system at work. Before the Gustard showed up, the NFB7.77 was in my home stereo for several months. I've had an Audio-GD R-28, NFB11.38, and now the NFB7.77. I can confirm that the R-28 (an R2R implementation) was a huge disappointment. It was interesting, but not really in a good way. I quickly offloaded it. I sold it because I thought the measly NFB11.38 was better. I liked the NFB11.38 (ES8030PRO) so much that I tracked down a used NFB7.77. I used that for awhile in my stereo until the Gustard X26 came out. The Gustard was an impulse buy. I like the fact that it is physically smaller than the Audio-GD with similar topology. Prior to these, I ran a Holo Spring KTE L3 with Singxer SU-1 KTE. My home stereo is for movies, TV, and music. The R2R sound of the Holo Spring is a bit too relaxed for movies and TV. I always felt like I was missing some detail. I believe the more detailed sound of the ES9038PRO is a better compromise to enable crystal clear TV & Movies and still very enjoyable music reproduction. I also prefer my music to be extremely detailed with laser focused sound stage & natural tonality. My reference for this sound is a Nagra HD DAC X.

My home stereo is

• Focal Electra 1028 Be & Rel Stentor II subwoofer (very rarely used)
• Sony TA-A1ES integrated amp
• Gustard DAC-X26
• I2S from Singxer SU-1 KTE & RPI running Volumio
• Nvidia Shield TV USB Audio or Optical from TV

Office headphone setup is

• Focal Clear
• Monoprice Liquid Platinum
• Audio-GD NFB7.77 USB input

I expected to use USB from my Nvidia Shield TV for movies & TV, and I2S from my Singxer SU-1 for music listening. This did not work out. The USB input to the Gustard generates an insane amount of latency. It's more than the 1000ms that my Nvidia Shield can compensate for. I would guess that it might be as much as 2000ms. For this reason I have to run optical audio out of my Samsung TV One Connect box. I have found no GPLL or filter setting which makes USB feasible for TV & movies. The latency is unusable. There appears to be no latency when using the optical input. This is strange. More investigation is required. This might not affect most people but I need a DAC that can interface with separate sources for TV/Movies and music.

The Gustard I2S input from the Singxer SU-1 KTE is excellent. The USB input direct from the Raspberry Pi is very good as well. The Gustard USB is without a doubt better than what is in the Audio-GD NFB7.77. My AGD has the Amanero Combo 384 USB input, but it does not have the isolator & separate power supply that AGD offers today. I did buy the isolator & power supply upgrade kit, but I ended up not installing it. It's a huge PITA. The I2S input on the AGD is what I would use for music anyway. The USB is good enough for TV and movies without any latency issues. I am extremely excited for an Auralic Aries to arrive today. I will see how that goes with SPDIF and AES directly to the Gustard. I expect there to be no difference feeding USB from the Auralic into the Singxer but I will obviously do that too.

The Gustard Gentle filter sucks. It just sounds like crap. The Composite filter is pretty OK. When the unit was new, the Vivid filter sounded like a compression filter which ruined dynamic range. Maybe my ears have changed, or maybe it's placebo, but the Vivid filter sounds a lot better now. I do not know if I like the composite or vivid filter better yet. I have never used CDT Opt because that name seems nonsensical and I do not understand it. But since VTory seems to like it, I will check it out.

The Gustard DAC-X26 has a very nice chassis and feels super robust. It runs very hot. Much hotter than my NFB7.77. This is easy for me to deal with but it is worth noting. Due to the small size of the unit you may be tempted to tuck it away into a tight location, but I advise against that. The power button on my unit is very crooked. This is annoying because the rest of the unit is so well made. This is the lack of attention to detail that I expect from Chinese goods. I will forgive it because the execution throughout the rest of the unit is well done and worthy of the price tag.

 

Pictures first, report compared with my Gungnir MB A1 will follow. The last one shows how thick the side panel is.

 

Gungnir MB A1 USB5 vs. Gustard DAC-X26.

Short version: If I had $1000-ish cash and wanted to upgrade from my Modi MB, I wouldn’t regret going with either one. But they sound different. My Gungnir MB A1 is the DAC that made me believe that digital music can sound very close to vinyl in 2016. The Gustard X26 is the DAC that made me believe that a D/S DAC can have clean and detailed sound and while being enjoyable in 2019. After two days of brief listening, I prefer to use the Gungnir MB for live recordings, and the X26 for most orchestra and some studio recordings. My Gungnir MB is emotional and vinyl-like, while the Gustard X26 is delicate, detailed, and forgiving.

tl;dl version.

First is the gear background. I only tested both DACs with aco-ax SPDIF input. I used a BlueJeanCable "5 foot BJC High-flex 3G/6G HD SDI patch cable (Belden 1505F), BNC to BNC" between my Pi2design 502DAC (streamer) and Gungnir. I used a BNC - RCA adapter then a "Straight Wire 75 ohm digital cable" between the 502DAC and the X26. Both DACs used "WORLDS BEST CABLES – using Mogami 2549 wire and Amphenol ACPR Die-Cast, Gold Plated RCA Connectors" to connect to a Schiit Saga, then a Vidar. The speakers were a pair B&W Nautilus 803, bi-wired. The speaker cables were a pair which are a little bit clinical and grainy compared to my Kimber 8TC. I used it for convenience as I have been switching speakers recently. These cables also served as a kind of a "brutal" test. I did not try the different filters on the X26. I used the "gentle" filter for this report. The X26 was 2-3 db louder than my Gungnir MB. I volume matched using a Radio shack SPL meter with a 1Khz test tone (both 71 db at my listening position). I like "tubey" and warm sounds in general. When I feel uncertain, I play vinyl to readjust my preferences.

The unit I tested is in silver finish. It is more silver than the pictures in the previous posts. Another thing to note is that this unit ran very, very warm for a DAC - think of it as being as warm as a class-A headphone-amp or a tube headamp. The heat is likely why the unit uses very thick aluminum panels on all the sides.

My first standalone DAC was a Rega DAC using two Wolfson 8842 Delta/Sigma (D/S) chips. I sold it because I wanted more detail. Then I got a Sabre 9018 DAC with the factory "high-end" clock upgrade and capacitor upgrades, which gave me headache after 15 minutes. Later I purchased the Gungnir MB and I found the sound that I have been looking for. In the last three years, being a Schiit MB DAC fan, I had listened to several AKM based DACs. Those AKM DACs sounded too digitally-processed. Every second I knew that I was listening to a Velvet sound DAC chip, but not actual music. It wasn’t until last Saturday, when I first heard the X26, than I thought: ok, the D/S DAC is not hopeless. When the right D/S chip is done right, it can be very good and enjoyable too!

Ok. Sound. Vtory and others had full scale reviews, so I won't repeat. I just wanted to share some bullet points here. The X26 is very detailed, which is one of the strengths of the Sabre pro line. The 9038 is also like the 9018, it digs deep into the details like the Shure V15 cart to vinyl. The plankton on the X26 is better than Gungnir MB. The feeling of ambiance is palpable. The sound stage is at the Gungnir distance but wider though thinner. The image is also a little better. The Gungnir MB did a very good job regarding on instrument separation. I felt even more space between the instruments after my Gungnir MB firmware update, but the Gustard X26 makes the instrument separation even better, which is obvious in jazz recordings. The detail is there but it is not intrusive. When I go back to Gungnir MB, I can still hear most of those details, but the X26 made those details easy to be picked out. The X26's bass is decent and has great texture. But the Gungnir MB has more bass extension. So far, it’s all good stuff. But when I passed the common "new gear excitement", I realized that the Gustard X26 is still in line with the modern D/S sound signature, which includes a very black background (very low noise level). It shows you how good and detailed your 44.1K digital music can be. I was astonished many times when I realized that my downstream pre-amp, amp, and speakers are detailed enough. I just didn't know it before. Now back to my Gungnir MB. It has very little haze even after the recent factory firmware update. But the MB still wins at realness and emotional involvement.

In the end. I found myself wanted to listen to more classical orchestra recordings when using the Gustard X26, as the details and the pin-point imaging let me feel the instruments on the stage more vividly. I still prefer using the Gungnir MB for human voice, piano, and live recordings, as the Schiit Gungnir does overtone better. Don't get me wrong, the Gustard X26 is pleasing for all kinds of music, it has never been fatiguing. I easily forgot that I was testing a DAC and was instead enjoying the moment. The Schiit Gungnir MB could do better on some recordings. But the Gungnir is not as forgiving as the Gustard X26. If I were forced to choose two words to describe the X26, I would say: black, detailed, and forgiving (oops, there’s three words). If I had to choose only one word, it would be "forgiving".

Final notes. I haven't explored the extreme potential of either unit. The Gustard could sound better when connected with the HDMI I2S input. When used with balanced XLR analog outputs, the Gungnir MB is reported to sound better as the RCA circuit has a summer circuit that can blur the sound a bit. Happy listening.

 

Well written. Good job! Except your assessment of gumb1 as "unforgiving" -- which I have never felt during my days with it -- I can easily agree with almost all your statements.

Nicely put. I am happy that it's not only me to hear this dac like this.

And couldn't agree more. With convert-2, I found I spent most of my head time with dance pop and electronics. X26 makes me to enjoy big band jazz and orchestral pieces more.

 

I think I did not describe this very well. I meant that when compared the Gungnir Multibit and the X26, the later was a little bit more forgiving to bad recordings. I felt that there is a very fine line between forgiving and disengaging. Those 9038 DACs (both the X26 and an cheaper one I heard) sounded very relaxing (forgiving) to me. But I felt that the 9038 DACs were also less involving then the Gungnir Multibit. The X26 still has the D/S DAC's digital sound signature, it's strength is that it doesn't sound fatiguing. But it is just a little less involving while it is little more forgiving than Gungnir Multibit. I'd try this: Gungnir Multibit can make me sweat on some CDs (like the live music does), the X26 can make me sit back, close my eyes, and simply relax.