When it comes to DACs, many people have the misconception that the actual chip used for D/A conversion is the only thing that really matters. That makes about as much sense as saying the drivers in a speaker are all-important, and the crossover/enclosure don’t matter much. When we put it in those terms it sounds like nonsense. Yet this thinking persists in the realm of D/A conversion. Experience has shown that you can have an excellent sounding DAC using a fairly basic DAC chip. Likewise, you can have a DAC that sounds unimpressive, even though it may use the latest/greatest chip (or even several of those chips). The total overall design is much more important than each individual piece.
With that being established, I do have to acknowledge what seems to be a superior DAC chip compared to many others on the market. Unless you are out of the loop, you have likely heard of the ES9018 Sabre Reference, the top DAC chip from ESS Technologies. Call it coincidence, but I notice that a very high number of products using this chip have gone on to become extremely well regarded. Examples: Wyred 4 Sound DAC 1/DAC 2, Eastern Electric MiniMax, and Weiss DAC202. In my personal experience, the Anedio D1 and Resonessence Labs Invicta are two of the best DACs I’ve experienced regardless of price, and both use the ES9018. Again, I certainly don’t think the chip itself is the only component involved in making these sound great. But it may allow them to reach a higher level than they otherwise could if they used a competing chip from Wolfson, Texas Instruments, etc. This is just an observation and I won’t pretend it is the only school of thought.
So we’ve got a high-tech new DAC chip, which appears in various highly-regarded DACs. Just choose one and you’re all set, correct? Unfortunately, nothing in audio is universal. Some people have complained that the “Sabre sound” can be too analytical, perhaps sounding a little off in the upper mids/highs. For that reason, some people have stayed away from the ES9018 based models. This especially applies to folks who love the stereotypically smooth “tube sound”. I don’t personally have this issue but there are many Sabre based DACs I have yet to hear.
This brings me to the topic of this review: the Yulong Audio Sabre D18 DAC. This is the first Sabre based DAC from Yulong, and they must have liked the ESS chip so much that they included “Sabre” in the name. This is by far the most ambitious and expensive design that Yulong has yet to produce under their own brand (they also do OEM work so they may have built higher end gear for other companies, I don’t know). This was meant as an all out statement product, while keeping a fairly low price – at $699 it is the most affordable DAC on the market which uses the ES9018, by a considerable margin.
I’ve reviewed several of the prior designs from Yulong in the past, including the D100 and U100 DACs and the A100 headphone amp. All have impressed me at their price points and even beyond. So I was curious how the flagship D18 would stack up against some of my favorite DACs. I’m happy to say that it holds up well – while not beating my reference units, the D18 brings something unique to the table that I have not yet heard from an ES9018 design.
The D18 is a fairly compact unit, using what initially appears to be an upgraded version of the D100/A100 enclosure, but is actually a completely new design. As such it should be able to fit into nearly any system: it won’t look too big next to a small headphone amp, nor will it look too small next to a full size amp. Weight is listed as just over 5 pounds but subjectively feels like more than that.
The main focus here is minimalism – the front panel has a power button and a source selection knob, and that’s it. The rear panel features RCA and XLR outputs, and an array of three inputs – coaxial, toslink, and AES/EBU. That’s right, no USB input. Apparently Yulong (the designer) feels that since there is already a plethora of competent USB to SPDIF converters on the market at reasonable prices, there is no need for him to reinvent the wheel. Those that need USB can choose from one of the many devices out there, while those that don’t are freed from paying extra for a function they won’t use. Obviously this will be somewhat of a polarizing choice – some will find it refreshing, others absurd. I do recognize that it is strange to see a “flagship” digital product without USB in the year 2012. On the other hand, the $699 price leaves plenty of room in the budget for a USB to SPDIF converter, while still coming in at less than most competing Sabre based units. Besides, this strategy isn’t unheard of - the well regarded Metrum Octave does the same, and Schiit Audio offers their Bifrost sans USB for a savings of $100.
Internally, the D18 bears a vague resemblance to the D100 in terms of general layout. Both use the same toroidal power transformer from Canadian firm Plitron. Both have a similar size PCB, and I see some shared components like the cable and fuse harware. Aside from those, the D18 appears to be improved in every area - better regulation and stabilization in the power supply, superior design and components in the output stage, higher quality system clock, and more advanced overall layout. And let’s not forget the ES9018 DAC, a ~$60 chip as compared to the ~$7 AD1955 used in the D100. Again, the chip (and the price) won’t tell the whole story, but it is worth mentioning.
The D18 is supposedly capable of accepting 24-bit/192kHz signals on all 3 inputs. I say “supposedly” because my reference gear tops out at 24/192 over coaxial or AES/EBU, and 24/96 over toslink. I suspect that the limitation of 96kHz over toslink is irrelevant since very few devices (if any) can actually achieve the theoretical maximum of 192kHz. The main point is that the D18 will accept basically any stereo signal you can throw at it short of DSD audio, which is only supported by a handful of devices. The ES9018 is actually capable of handling up to 32-bit/500kHz signals, but the SPDIF standards do not allow anything higher than 24/192. While some companies are using USB to handle higher rates, there isn’t really much point since the vast majority of music is 24/96 or lower, with the bulk of it being “lowly” 16/44.1.
The output circuit uses a pair of Analog Devices AD8620 opamps for I/V conversion, and a pair of AD797 opamps for low-pass filtering. I’m told the output stage appears very similar to the IVY module from Twisted Pear Audio, but I have no experience with that so I can’t confirm. In any case, it is far more elaborate than Yulong’s own D100. I’ll let the pictures tell the rest of the story. One novel feature that you may notice is something looking like a cable, connecting one area of the PCB to another. This is called an SMA interface, and it is used as a bridge between the input section and the ES9018 DAC. While it would normally be carried along through the PCB, Yulong says the SMA method does a better job of keeping this critical signal isolated from noise and interference. Once delivered to the DAC, things proceed in a more traditional manner. Note the custom made low phase noise crystal oscillator parked directly adjacent to the ES9018.
ESS 9018 and system clock
Analog output stage
Since Grant Fidelity is now the exclusive North American distributor of Yulong products, I asked them to assist me in getting some information. I don’t speak any Chinese, so Rachel at Grant Fidelity was very helpful. What follows is her translation of the questions and answers, with some editing for clarity.
1) Can you explain the why the RCA outputs sound slightly different than the XLR? Is that something you went for on purpose, or did it just happen that way?
A: The RCA and XLR output sound difference is designed on purpose to cater different headphones and music styles. It is also for better matching with the coming Yulong Sabre A18 headphone amp.
2) Can you give rough jitter estimates? I know everyone measures differently, but some idea would be good to have.
A: With 1KHz signal, Yulong Sabre D-18 has jitter measurement better than 1ps RMS. Sabre D-18 uses very high end reference clocking at phase noise -130dB. We consider the measurement to be outstanding even by high end standards.
3) Can you talk about the matching A18 headphone amp - any comments on the design or release date?
4) What "sound" were you going for with the D18, and do you feel you achieved it?
A: Yulong Sabre A18 is a high end Class A headphone amp with K170/J74 JFET input stage. We consider JFET circuits to have better sound quality than vacuum tube circuits for input stages, and it will deliver rich smooth sound. The unit has low total harmonic distortion and wide frequency response. It comes with RCA and XLR inputs (real fully balanced circuit design) so that you can match it with other components. It is designed as a perfect match to Sabre D18 to show you the absolute full potential of what the Sabre D18 can achieve. I designed the Sabre A-18 for my pursuit of high end sound quality as an audiophile - in the design process, no cost objection was considered. I was after a rich, deep, smooth, detailed and dynamic sound and I personally think that A18 has achieved my design goal. It has low output impedance, sounds very musical, and matches well with a majority of headphones on the market. Launch time should be next 3-6 weeks, please check with our North American agent Grant Fidelity for official exact launch date.
I cannot give more specific description of the Sabre A18/D18 sound. Listening is a subjective matter and people may have different opinions on how to interpret the sound they have heard. I will leave it to customers to provide their own description and unbiased feedback.
5) Was there any area where you had to "cut corners" with the D18 in order to keep the price low? Or is this a true statement product with nothing held back?
A: I wouldn't say 'cut corners' but in production we need to have an overall cost control in order to provide the final product at a suitable price. Keep in mind that production cost in China is much lower than in western countries and our worldwide universal retail price also means much less mark-up in the pricing structure. People in USA or Europe are paying the same price as if you personally come to China to buy from a dealer here. Sabre D18 and Sabre A18 are both of very high end sound but at much less price point compared to their American or European counterparts. I personally refer to this as 'value' - we provide high 'value' products - not just high end or high price. It's the ratio of price vs. quality that matters in the end to consumers. So it is our statement product with nothing held back in design, but our production/distribution model is ultra lean and efficient. The buyer's money has been most effectively spent on sound, not on reselling cost.
6) Did you find the ES9018 difficult to work with? I've heard that from some designers.
A: ES9018 is quite different from any other DAC chip on the market so there is a learning curve involved for best utilizing the chip in a final product. We spent substantial time to develop the Sabre D-18. We are planning to develop more products with the ES9018 chip down the road.
I’m not sure exactly what to make of the 1ps RMS jitter measurement. I know that CEntrance makes the same claim with their DACmini products, but I’m not sure either company is testing as thoroughly as the folks at Audiophilleo or Anedio. Both of those teams have extensive measurements available for the end user, and both have mentioned how difficult it is to get an accurate number when you are dealing with such low levels. In any case I think it is still safe to say that the D18 does an excellent job, and even if they are off by a factor of 10 it would still be a very good result.
Also worth checking out is THIS post on the Grant Fidelity website. It documents their trip to Yulong headquarters and shows some pictures of Yulong’s work area. This guy is clearly “one of us”, owning the Sennheiser HD800 and beyerdynamic T1 among many others. He is also clearly serious about establishing his company based on quality rather than price alone. He is very heavily invested in measurement equipment, and would rather the company grow slowly in size and reputation rather than compromising to speed up the process.
The D18 uses a nice thick aluminum material for the enclosure. Again, see the pictures – it’s thicker than the D100/A100 units. The source selection knob feels like solid aluminum too, and doesn’t come off without removing a screw, thus addressing a complaint some people had about the D100. It moves with a very reassuring “click” from one input to the next, and a corresponding LED lets you know at a glance which source is selected. Even the front panel power button has a confident feel to it when pressed. Overall it feels more like a flagship product than the D100 did (though the D100 was pretty nice as well).
The D18 comes in the usual Yulong packaging. It’s efficient if not particularly extravagant, and should keep the item protected well enough during its journey overseas. There really isn’t anything to this device that would need explaining in the manual, so they throw in some measurement printouts.
This is the associated equipment I used to evaluate the D18.
Source: Lexicon RT20 universal disc player, JF Digital HDM-03S digital audio server, NAD C 446 network audio player
Amplification: Violectric V200, Analog Design Labs Svetlana 2, Yulong A100, Matrix M-Stage
Headphones: Heir Audio 8.A, Unique Melody Merlin, Lawton Audio LA7000, Sennheiser HD700 prototype, Earproof Atom, Ultrasone Edition 8, Audio Technica W1000x
Cabling was by Signal Cable, music played ranged from CDs and 16/44.1 FLAC files all the way up to 24/192 releases. I let the D18 burn in for over a week solid prior to doing any listening (not that I think it is required but I was busy).
This is not my picture, but someone on Erji.net has both the D18 and the
Weiss DAC202, and says the D18 is a very good product. The intricacies
of the comparison were lost in translation.
Some of my listening and comparison gear
The Yulong D18 is an excellent sounding DAC. Nothing stood out to me upon first listen, which I take as a good thing – if I notice a gimmick like spotlit highs or artificially large soundstage, I might be impressed the first hour, but that sort of thing is sure to get old in the long term. So in my opinion all high quality DACs should sound very similarly excellent in general. Once you get into some heavy listening, you’ll start to discover the intricacies that make each one unique, but I don’t feel the comparison to be as easy as some people make it out to be.
First off, the D18 is on a high enough level that it doesn’t really require a step by step breakdown of lows, mids, highs, etc. As befitting a truly “high-end” device (in performance rather than price) it handles everything very well. Bass drums go staggeringly deep, brass has the proper bite, vocals are clear and lifelike. I would expect nothing less than this considering the fact that the lower priced D100 pretty much already had these things nailed. So in the spirit of making my reviews more readable, I’m going to assume that the reader has heard a top class DAC at some point in the past, and knows how it goes once you climb the ladder high enough.
Getting into those intricacies that I mentioned earlier – the character of the D18 is best captured by the word “smooth”. This is absolutely a DAC designed for musical enjoyment rather than sonic analysis. It gives the perception of warmth in the mids that deviates from the more clinical D100 sound, though I doubt it is the sort of thing that would come out in a frequency response measurement. There’s just some more energy there, giving a measure of extra weight or authority to the sound. But it never goes too far by sounding overly thick or slow.
The smoothness also extends to the highs – it brings to mind the difference between a stereotypical solid state amp, compared to the stereotypical tube amp. The tube sound is not necessarily any darker or more rolled off, but still manages to take some of the edge off, facilitating long listening sessions and an overall lack of listener fatigue. Despite this smoothness, I feel that the D18 is still a very detailed DAC. There’s nothing missing per se – it just sounds a little smoother, a little cleaner than it might with something like a Benchmark DAC-1. Add in the wonderfully creamy mids and the prodigious low frequency response, and I can see how this DAC could be described as being on the warm/dark side. I’m not willing to call it dark, or at least not in the same sense that some headphones are dark, but I will concede to calling it warm. Once again - I certainly don’t mean to imply that there are missing details here; the DAC is very precise overall and extracts high levels of musical information. But it has a tendency to make things sound a little better than they really should. It’s not exactly forgiving in the true sense of the word: mp3 files and poor recordings will still show their warts. But it does give some flaws the benefit of the doubt, where something like the Benchmark unit would have no tolerance whatsoever.
Another excellent attribute of the D18 is the ease with which it serves up the music. It feels like everything just flows in an organic manner. Instead of lows/mids/highs you have a singularly cohesive experience from top to bottom, with no separation. Yet if you want to focus on any single aspect, you certainly can. Note that this is different from the idea of “instrument separation”. Though falling just short of my reference units, the D18 does allow for easy tracking of each individual performer, in a way that few (if any) sub-$1k DACs can match.
One interesting feature that could be considered a quirk or a benefit depending on your point of view: the D18 has a different sound when using RCA outputs instead of XLR. This being a true balanced DAC, the single ended output is generated from the balanced signal rather than the other way around like many other designs. As Yulong mentioned, he wanted to give people options, especially those who will be pairing their D18 with the upcoming flagship A18 headphone amp. Since that amp has both XLR and RCA inputs and both single ended and balanced headphone outs, it would be like having two different sounds in one setup. As far as the D18 goes – the RCA output seems to me like it has more of the traditional ES9018 “Hi-Fi” type sound. It’s still smoother on top than most, but lacks some of the warmth and body that balanced mode brings. Using the XLR output gives that really analog type feel, which is what I used most during this review. But don’t misunderstand – the RCA outputs still sound excellent, and with the right amp may actually be more appropriate than the XLR option. Keep in mind that you can easily use an XLR to RCA cable to extract the balanced flavor from this DAC, even if your amp is limited to RCA inputs. Also keep in mind that many amps will have differences of their own when using RCA inputs versus XLR – so it really comes down to mixing and matching to find the best synergy. The overall character of the DAC remains, so don’t expect night and day differences. But there is enough to give you some variety.
An early version of the D18 and matching A18 amp
The D100 is an excellent DAC in my opinion. It is clean, transparent, detailed, and very representative of the skill Yulong has as a designer. I still gladly recommend it for people who want a great buy in the sub-$500 price range. It’s hard to talk about the superiority of the D18 without making it seem like the D100 is deficient, which is simply not the case. The fact that the D18 is superior in many ways does not take away from the accomplishment of the D100.
One of the biggest improvements offered by the D18 is the immersive, three dimensional soundstage. The D100 was certainly no slouch in that area, but the D18 takes it to that next level where only a small number of DACs can go. It was a modest gain when using most headphones, but when I tried a model with excellent soundstage performance such as the HD800 or the UM Merlin, I could discern a notable improvement. I heard similar improvement when listening through a speaker rig. There is just more “space” in the performance, both in terms of size as well as capturing the acoustics of the performance venue. When playing very good recordings from Mapleshade, Reference Recordings, MA Recordings, and other top notch labels, the D18 transports you “there” to a higher degree than the D100.
The other big difference is with how the top end is handled. Some people have found the D100 to be a bit bright when paired with certain gear. The D18 will not have that issue. It presents similar levels of detail, but in a completely smooth and effortless way. Is the D100 more accurate? Maybe sometimes it is, depending on your definition of the word. Is accuracy defined as reproducing the recording exactly, flaws and all? Or does it mean making a sound that seems lifelike, the way a real instrument is likely to sound, even if the recording may not reflect that exactly? Whatever your take on it, I suspect that the D100 would probably make for a better tool in a home studio situation, where you really want to hear those flaws. The D18 would be a better match for most “audiophile” type listening scenarios, where the listener is more concerned with making their music sound as enjoyable as possible. On the whole I do think the D18 is very much the superior machine and easily justifies the price increase.
I have to preface this comparison by saying that I don’t have the Quattro DAC at present – I sent it around to some HeadFi members for audition, and it’s still making the rounds. But based on my listening notes and recollection, these are very different sounding DACs.
The Quattro is significantly more lit up in the upper mids and highs. Its hardware is very similar to the Benchmark DAC-1 and the sound generally reflects that, so the D18 has a completely different take on things. The strong points of the Quattro come close to matching the D18 – soundstage is nearly as wide, though a bit less convincing, details come through very clearly, and low frequencies are well done if not quite as controlled. The mids are leaner and the highs seem to be more grainy, though only in context of being compared to a nicer DAC. The bottom line for me is that the D18, though identically priced at $699, is indeed the nicer sounding DAC. There’s a catch though – the D18 doesn’t have nearly the functionality that the Quattro does. The Matrix unit offers the same three inputs as the D18, and gives the same choice of single ended or balanced output. While the D18 stops there, the Matrix offers a hi-res USB input, a respectable integrated headphone amp, an analog input, volume control for pre-amp functionality, and a remote control. So while the D18 is strictly a high-end DAC for non-USB sources, the Quattro is a multi-function all in one DAC/pre-amp/headphone amp. And it sounds pretty good too, competing well with the Yulong D100. So the two devices have very different target markets.
The Anedio D1 is my reference unit, and in my experience has only been surpassed by the much more expensive Resonessence Labs Invicta. It is soon to be replaced by the updated D2 model (which I will be getting when it is released), but for now the D1 is the DAC to beat as far as I’m concerned.
At $1270, the D1 is nearly double the price of the D18. But they both share the same ES9018 at their core, so I figure it is a good comparison. The D1 has more features like USB in, an excellent headphone amp, remote control, and pre-amp functionality. But the D18 has advantages of it own – AES/EBU input and balanced XLR outputs.
Soundwise, these two share many similarities but also diverge in overall character. I’ll say right now that the D1 is still in a higher class – but that doesn’t surprise me, since the D1 is one of the best DACs I’ve ever heard at any price. The D18 is impressive in that it can approach the performance of the Anedio in some ways, for a significantly smaller investment.
The similarities are found in the ability of both units to convey inner detail. Both are spectacularly good in this regard. Both also have an excellent command of the presentation: whether the track is a solo vocalist or a complex orchestra piece, both units handle the performance with ease, allowing the listener to easily focus on whichever individual aspect they want (or just enjoy the presentation as a whole).
The differences: the D1 is a more linear presentation from top to bottom, with the main deviation being notable in the highs. Using RCA or XLR out of the D18, it always seems more smoothed out on top, with a subtle masking of imperfections. At times this is a desirable effect – many of my favorite albums are low budget, relatively poor recordings. Other times it is deleterious – the D1 takes good recordings, especially hi-res stuff, to higher levels.
Obviously the DAC is not the only contributor to the final sound, and you need to consider the entire chain including the headphones. If I want to listen to an album I enjoy, but that I know isn’t the best recording out there, I often reach for my setup consisting of the D18, an Analog Design Labs tube amp, and something like the Kenwood KH-K1000. This is a highly resolving yet musically satisfying setup that maybe isn’t the last word in technical ability but just hits my ear the right way. Examples would be when playing a Telarc release like “Some of my Best Friends Are… The Trumpet Players” by The Ray Brown Trio. Plugging the fairly smooth Kenwoods straight into the excellent D1 headphone amp, the result is still very much on the bright side. That’s not the fault of the Anedio; it’s simply there on the disc. But I might not necessarily be in the mood to deal with that for long term listening. Another example is the self titled release from Sister Sparrow and the Dirty Birds. I love the music, but the recording has a sort of dry, closed up feel to it, which is the exact opposite of the vibe you get when seeing them perform live. The D18/AD Labs/K1000 combo injects a welcome amount of warmth and life to the presentation; it still doesn’t capture the scale and energy of the live performance, but it engages the listener more than something like the Anedio/HD800 combo would. Again it comes back to the question of which version “accuracy” you are chasing.
I’ve said multiple times that the source is not the best place for any type of coloration, however small it may be. And I stand by that. If you have the choice, your best bet is to get an utterly transparent and neutral DAC. Can you find something which has the resolution, dynamics, soundstage precision, and sheer grace of the D18, for anywhere close to the $699 asking price? In my experience the answer is no.
I applaud Yulong, both the man himself and the company in general, for bringing the D18 to market. It would have likely been easier to make a fairly neutral device that built on the foundation set by the D100, and I bet it would have sold well enough. But instead they found what appeared to be an empty spot in the market, and filled it. I’ve long heard people complain about Sabre-based DACs sounding really great in most aspects, yet being somewhat lacking in musicality. I believe the D18 may be the perfect answer for those folks. In many ways it offers reference level quality, yet at times it departs from that in the interest of listenability.
The D18 will obviously not be for everyone. The person who uses something like a Benchmark DAC with a Gilmore GS-1 and HD-800, and wants to upgrade their DAC while maintaining a similar overall sound, should probably not be looking at the D18 to achieve that goal. The person who prefers a “tube sound”, likes vinyl, or just has a hard time enjoying the hyper-detailed sound that many of the best DACs provide, should absolutely give the D18 a try. Grant Fidelity offers a 30-day in home trial, so it would merely cost the price of return shipping to try the unit in your system. I suspect in many cases that return shipping would not be necessary.
Edited by project86 - 2/5/12 at 12:13pm