|Posted by Amine Slimani on October 12, 2010 at 7:10 PM|
Before getting into the second part of this review, I wanted to take a moment to describe the effect of a small “tweak” that can applied to many USB converters, the USB2ISO or ADUM isolator.
What this little dongle does is that it provides galvanic isolation between the computer and whatever converter is connected through USB.
USB2ISO isolator (galvanic USB isolation):
Many USB converters have some sort of transformer at their spdif output stage but most of them have zero to little isolation on their usb input.
While a single proper transformer between the usb converter and the DAC is enough, in theory, to protect the converter from the noise generated by the computer, the USB noise can, in fact, affect considerably the performance of the USB converter.
Also, it has been reported that some poorly made pulse transformers may not block all the noise coming through the USB line.
Anyway, since the ADUM/USB2ISO isolator was priced at a relatively affordable 35 euros price, I decided to give it a try a try and check if the reports of people noticing big improvements through that little device held any truth.
During the 2 months I had the device I tried it mainly on 3 converters the Musiland 01 USD, Teralink-X2 and the A-GD DI. (Note: not all devices are compatible with the device, and for instance, the Hiface is not).
The effects noted on the converters were rather similar, but some converters were more affected than other.
Some of the most noticeable changes were a removal of digital grunge in the highs, and an improvement in clarity,imaging and soundstaging.
On the Musiland the effect was so dramatic that it made very "listenable". Meanwhile, the effect on the Digital Interface with the external power supply was barely noticeable.
Below is how I would rate the converters starting with the most affected to the least affected by the isolator:
- Musiland (most affected)
- Direct USB input of my DAC (similar to the Musiland)
- Teralink-X2 with external power supply
- Digital Interface
- Digital Interface with external power supply (barely noticeable)
So while I have found that the USB2ISO can improve the sound of most entry to mid level converters I had on hand, I found that different converters reacted differently. The better the power supply (and filtration) of the converter, the less improvement there was with that little USB device. It is possible that after a certain point (i.e. high-end converters) the USB2ISO converter might not provide any improvement and might as well degrade the performance by affecting the latency.
I personally found that the effect was well worth with all the devices I tried it with save for the Digital Interface with the external power supply. For that specific configuration, the effect is very subtle and I can even imagine people preferring the Digital Interface without the USB Isolator.
Keep in mind that the effect of the USB isolator will depend on the quality of the PS of the computer being used. When doing any critical listening, I either use my laptop running on battery or plugged to a power filter (with the battery removed). So your mileage may vary depending on your specific configuration.
The USB to SPDIF shoot-out Review - Part 2 : Teralink-X2 vs. Jkeny's modified Hiface vs. Audio-gd digital interface
I got the Teralink-X2 while my Hiface was broken (due to the excessive weight of my Oyaide Digital cable was applying to its usb solder) and was being repaired by m2tech.
The Teralink-X2 was a nice improvement over the X1 by adding 24/96 capability as well as improving the sound quality. However, like other Tenor based converters, it lacked 88.2K support.
Both models (X1 and X2) seemed to share the same warm and smooth sound, but the X2 was the most resolving of the two.
While, its sound improved by adding an external power supply and the USB isolation, I felt that it was always lacking in realism of timber. The stock Hiface, while more lightly balanced, seemed to be (comparatively) better at preserving the natural timber of instruments and voices.
By applying the tweaks mentioned above, the subjective noise floor seemed to drop, the soundstage was even bigger and more defined, the level of details increased, and the bass became more articulate.
But those improvements didn't outweight what I had previously noticed about the X2, it always seemed to me there was something off about the sound: mainly a slight "out of focus" of imaging and timbers.
I have to admit that if I didn’t have a superior converter such as the battery powered Hiface; I wouldn’t probably have been able to pick up on those faults.
Overall, I used the Teralink-X2 the past few months as a back up unit whenever the Hiface batteries were depleted and needed recharging.
I would recommend the X2 for someone starting to build a system and needing a 24/96 usb transport for a very affordable amount. I personally felt the stock Hiface was better than theTeralink-X2 and required less hassle such as external power supplies, usb cables, and USB isolators.
Side Note: Keep in mind though that when I am describing the stock Hiface, I am referring to the performance of the older models with 2 big clocks. At one point during the life cycle of the product, M2Tech used miniature clocks for the 44.1K multiples frequencies. Weirdly enough, it was about the same time new owners started complaining about how thin, bright and forward sounding the stock Hiface was sounding.
M2tech said the new miniature clock had the same specs as the big clocks. Though it seems that M2Tech has reverted once again to the bigger ones.
This is a non-issue for the battery powered Hiface since the direct battery power supply to the clock makes the quality of the clock less relevant. (The better clocks have better powersupply noise rejection capabilities than lower end ones).
Jkeny’s modified Hiface:
(Edit: The review below is extracted from the jkeny's Hiface review I wrote earlier this year)
Timber & Tonal balance:
While there is no shift in tonal balance per se, there are however differences between the stock Hiface and the modified one in the way they affect the DACs. (Edit: The more solid bass representation of the modified Hiface may change the perceived tonal balance into a more neutral one than the comparatively lighter sounding stock Hiface).
The most striking feature of the modded Hiface is a better extension at the frequency extremes: the bass extends deeper and the highs extend higher and in a cleaner fashion than the stock Hiface.
In comparison to each other, the sound of the modded Hiface is fuller. But this is different from the false warmth injected by some jittery converters. While the Teralink X2, for instance, is indeed warmer than the stock Hiface, it achieves so by adding the same extra warmth and haze to all the sounds, which reduces the overall resolution.
So what to expect from the modified Hiface? With well designed and neutral DACs, the sound gains in tonal density and realism. The instruments are more easily recognized by gaining in realism.
Another strength of the modded Hiface is the way it renders the specific timber of very close sounding instruments.
Listening for instance to the Concerto for 2 violins - Vivaldi, you can hear that Carmignola and Mullova are playing 2 different violins (a Stradivarius and a Guadagnini if my memory serves me well). The violins are not only beautifully rendered but they also have distinct tonal signatures. While some warmer converters can make this album tolerable to listen to, they rob the inner details of the instruments and can make everything sound the same.
Of course, this won’t be the case with every DAC, the better and the more transparent the DAC, the more subtleties you will be able to enjoy.
Soundstage & Imaging
Here, the modified Hiface achieves something really interesting. It can throw a huge soundstage (if the components downstream are up to the task) while retaining a superb imaging capability.
If we take the Teralink X2 for example, it can throw a pretty big soundstage but it lacks depth and it is fuzzy and blurry.
After listening for a while to the modified Hiface I understood that part of what the Teralink X2 was doing was to push the soundstage further back which is good for headphone listening and with entry level DACs. However, by doing so, It also blurs the soundstage layering and reduces its overall depth.
The stock Hiface throws a smaller soundstage in comparison to theTeralink X2 but it is a lot more defined a layered.
What the modified Hiface does is to throw a bigger soundstage than any other converter I have tried to date. To be more specific, the soundstaging is upfront like the stock Hiface but it has a tremendous depth which makes it seem a lot bigger overall.While listening to the modified Hiface, you just feel like having an open window into the representation.
I have previously described the stock Hiface as having a holographic imaging. Here the modified Hiface goes a little bit further. Thanks to its greater tonal density and to its more accurate timber, you get not only the holographic imaging of the performers, but you also get a greater sense of realism. On good recordings, you feel there are real persons breathing and performing in front of you.
From a technical point of view, the stock Hiface was already very good in that regard. But whatever, the modified Hiface is doing, it seems to be more convincing emotionally speaking.
Overall, the modified Hiface makes it a lot easier to mentally “reconstruct” the place where the recorded event took place. It just makes much more sense when listening to the modified Hiface.
I first thought that I wouldn’t have much to write about in this section of the review. When listening to the modified Hiface, everything was fine; It had huge macro-dynamics and beautifully rendered micro-dynamics but I assumed that it was coming from the rest of the chain. It was until I reverted back to the TeralinkX2 and Stock Hiface that I noticed the difference.
When trying new audio gear, we can get accustomed sometimes too quickly to its benefits and it is only until we revert back to the previous equipment that we realize how far we have come from.
Here, the modified Hiface was actually a pretty big step up in dynamics. In fact, reverting back to the stock Hiface (or worse to the Teralink X2), the sound become duller, with slowed transients.
Once again, since the modified Hiface is only a usb to spdif converter, the limiting factor in most situation will probably be the DAC itself. But in comparison to other converters, the modified Hiface gave the most dynamic results regardless of the DAC.
Hence, my guess is that is this characteristic will be audible in all DACs.
Transparency & Definition:
The Battery powered Hiface has an excellent analyzing capability. It can dig very deep in the recordings but renders the information in a very natural way. While the stock Hiface is slightly more upfront with the details, the modified Hiface has a lot more low level details. With the right associated DAC, you can get a very detailed and relaxed representation.
Personally, I have never heard in my system a converter as detailed and at the same time as "analog like" as the modified Hiface.
One interesting thing I also mentioned in my review about the dac19dsp is that with the modified Hiface it is the first time I can clearly hear a big improvement on 24/96 files.
Before that, I used to have a hard time distinguishing properly upsampled 16/44 data to 24/96 from the native 24/96 ones. In fact what I had realized is that most differences we hear going from 16/44 to 24/96 are due to the poor filtering at 44.1. Even with my entry level dacs (emu 0404 usb, audio-gd dac100, audio-gd FUN, Purepiper DAC A-1, Zero DAC), I can hear the differences between 16/44 and 24/96 but simply because their digital filters are relatively poor at 16/44.
With the modified Hiface associated to the dac19dsp, it was something else. I really heard an increase in resolution by going to 24/96.
Contrary to what I am used to, I have made very few musical examples simply because it is very dependent on the associated DAC.
The modified Hiface is a very transparent device. It doesn’t have a sonic signature of its own. As far as I could tell, the overall sonic signature will mostly depend on the associated equipment and the recording itself than on themodified Hiface itself.
For more details about musical examples, I invite you to read my review of the DAC19DSP for which I used the modified Hiface as a transport.
While the modified version that jkeny sent me is still a prototype, I can say that it by far the best converter I have listened to in my system. I was expecting a small improvement (since I considered the stock one to be excellent) but I was totally surprised by the level of performance of the modified Hiface. Its effect on the sound was not subtle at all.
Jkeny described the sound of the modified Hiface as similar to that of expensive analog and I couldn’t agree more: his modified Hiface makes the other converters sound broken.
Personally, I have already asked jkeny if I could send him my stock Hiface tohave a similar mod done to it.
Audio-gd Digital Interface:
Preliminary impressions (after 2 days):
So far I have only listened to it in the following configuration: USB2ISO (usb isolator) - Wireworld Ultraviolet USB cable - Digital Interface (w/ external Power supply) - OyaideDB-510 - DAC19 DSP. I didn’t have time to try any other configuration, so keep that in mind when reading my comments.
Below are some preliminary impressions:
- Excellent low level details retrieval (similar to the modified Hiface)
- Excellent resolution: 24 bit material is really perceived as superior to 16 bit equivalent (similar to the modified Hiface)
- Big holographic soundstage, excellent imaging, but lacks a little bit in depth incomparison with the modified Hiface
- Excellent top to bottom coherency, the timber and pitch of instruments sound very realistic (similar to the modified Hiface)
- Very smooth and distortion free treble (better than the modified Hiface)
- Excellent transient speed
- Very dynamic sounding (perhaps even better than the modified Hiface)
- Tonal Balance: a little soft sounding at 44.1K and neutral sounding at 96K
- Excellent upsampling algorithms (to 96K): No loss in the low level details or naturalness, the tonal balance improves (in my system) and images snap into focus.
Overall, at this point of burn-in, I feel that Jkeny’s modified Hiface performs at a slightly higher overall performance. The difference is also more noticeable at 44.1 and not so much at 96K.
However, when it is compared to entry level usb converters (such as the Teralink X2, Musiland), the Digital Interface is in another category.
Keep in mind that these are some very early impressions that may change after I get more time listening tovarious recordings with the DI.
My only complaint so far isthat it doesn’t accept 88.2K.
Since I listened to Jkeny’s modified Hiface, I have felt that most other converters sounded fake and fatiguing in long term listening. The Digital Interface is the only otherconverter I have tried so far that seems to be good enough that I don’t feel like I am missing the sound of the modified Hiface when listening to it.
More on the DI (after 1 week)
External Power Supply
My Digital Interface came with the external power supply. I did my trial on 2 systems:
System A: DI + Oyaide DB-510+ FUN + HD-650 (w/ stock cable)
System B: DI + Oyaide DB-510+ DAC19 DSP + C2 + Beyer T1s (w/ ALO upgrade cable)
On the less resolving system A, the addition of the external power supply was barely audible. I would have to spend a lot of time to hear any difference at all on that system.
On the more resolving system B, the addition of the external power supply was clearly audible, but the extent of the improvement was dependant on whether the external power supply was plugged straight to the wall or through my Bada power filter.
When plugged directly to the wall outlet (or even through the unfiltered socket of my Bada filter), there was a small increase in low level information and a slight tightening of the sound. However, it was nothing worth loosing sleep over.
On the other, hand when the external power supply was plugged in my power filter, the increase in resolution, frequency extension at the extremes, and soundstage depth were more noticeable.
On my reference system, when operated straight from USB, the DI was better than the Teralink-X2 (with external power supply) but not as good as the modified Hiface. When using the external power supply, and plugged to a (good) power filter, the performance was roughly on the same level as the battery powered Hiface, and even better on some parameters (as explained below).
When I plugged my DI on my XP based laptop, it was recognized as the Teralink-X2 since both devices use the same Tenor usb chip. I have since played the DI through Foobar with theTeralink-X2’s ASIO drivers. I haven’t felt compelled to try other configurations, and since it sounded good that way, I stuck with it for the remaining of the review, to keep everything constant.
Resolution and overall definition:
Like other high quality converters, the DI has 2 paradoxical qualities: on the one hand, there is plenty of resolution at low level listening, and on the other hand, it is also possible to raise the listening volume to extremely high levels without feeling any distortion or fatigue.
In my book, it means that we are dealing with true resolution and not fake resolution (that would be the result of mid treble brightness for example).
While the increase inresolution over entry level converters was noticeable through my HD-650, it was far more obvious on the ALO recabled Beyer T1s.
Listening to “Open Your Ears - Chesky/Head-fi” which I have bought in both 16/44 and 24/96 versions, you get a clear sense of increase in perceived resolution when listening to the 24/96 version. The sound in the higher resolution version is more open, less digital sounding, more detailed and “faster” (less transient distortions).
Upsampling the 16/44 files to 24/96 (using the DI built-in upsampling) does improves slightly the focus and gets a little bit closer to the native 24/96 files. This is the first time Ifind that an upsampler provides me with a clear improvement and not just a sideways shift with my dac19dsp. The excellent SoX upsampler didn’t sound as transparent but has the advantage of being very “tweakable” (you get to choose yourself the phase and passband settings).
Throughout that test CD, you get to understand what high resolution is about. Very little low level details that had been buried previously with entry level converters just reappear and make for a very convincing and lifelike representation.
Though, I have to note thatthe DI seems to be more convincing at preserving the resolution of the recordings at 48K and 96K than it does at 44.1K frequencies. I assume that the fact the Hiface has 2 clocks (one for 44.1K multiple and a second one for 48K) helps keeping the same level of performance regardless of the sample rate.
BTW, when I talk about low level details (for the DI vs. modified Hiface), I am talking about insanely low level details. Even the Teralink-X2 can render the ambiance of the recording venue and the little noises (the hands of a performer on a flute, people coughing, chair noises...). What the DI and the modified Hiface allow you to hear more clearly is the traffic noise outside a recording venue for example. So if your downstream components don’t have a vanishingly low subjective noise floor, those differences might be a non-issue.
Soundstage and Imaging:
Regarding the soundstage, the DI gives excellent results similar to what I have achieved with Jkeny’s modified Hiface. With a “superior” headphone that uses angled drivers such as the ALO Recabled Beyer T1, they both can throw a wide and believable soundstage in front of the listener. Unlike the somewhat big and diffused soundstage of the Teralink-X2 that is relatively flat and undefined, the DI opens a very transparent window in front of you.
The DI has superb imaging qualities: the higher perceived resolution and contrast ratio help you to picture very easily individual performers and instruments. On very fine recordings (mostly classical), you can guess easily how the performer was facing the microphone, or how a soloist was playing his violin.
Timber and tonal Balance:
Among the things thatseparate the DI from entry level converters is the “weight” of the representation. Most built-in USB inputs (I have tried) and low performance converters seem to be lightly balanced. As for the stock Hiface, while it preserves relatively well the timber of instruments, it could also be described as lightly balanced. The DI on the other hand, has a bass can reach very low and has a very fine texture. The representation is very “weighty” without it being bloated. To be more specific, and to give an example, the DI hits harder in the lows than the warm sounding Teralink- X2 but, at the same, the bass of the DI is more nuanced and articulate.
Like the Jkeny modified Hiface, the DI uses its superior low frequency to anchor the images into the soundstage and enhance the believability of the soundstage, which is a very fragile parameter on headphone listening.
I have read one or two people describing the DI as being dark sounding. Well, if you compare it to the Musiland, EMU0404 USB or (from memory) to the stock Hiface, it is definitely darker.
But I feel that the DI is not as dark as the other converters are being bright sounding. The top 2 converters I tried in my system (Jkeny’s Hiface and the DI) have somewhat a rather similar tonal balance in comparison to brighter and harsher sounding entry level converters.
As I have said in my initial impressions, the DI seems to have a very even tonal balance throughout the frequency spectrum, with no apparent aberrations. Its bass is very deep articulate, and musical. Its midrange is very transparent and the treble is very smooth and distortion free. In fact its treble reminded somehow of what you get when moving from an entry level sigma delta DAC to a good R2R DAC such as the dac19dsp. You get at first the sense that there is less treble but the more you listen, the more you realize is thatwhat you get is in fact less digital hash and distortion and more (truer) high frequency information.
On one of my reference tracks, Concerto for Two Violins from Vivaldi,Carmignola and Mullova are playing 2 different violins: a Stradivarius and a Guadagnini. With an entry level converter such as the Musiland, the violins sound unnatural. On the DI, similarily to Jkeny's Hiface, you not only get natural and realistic sounding violins, but you can very clearly hear the slight tonal differences between the Stradivarius and Guadagnini violins. Of course, you need a pretty resolving system to hear such distinctions.
Also, thanks to its superior low level details retrieval, you hear the two soloists breathing, moving and you can even guess how they were playing their violins and facing the microphones.
The naturalness of the representation, coupled to the superior low level information retrieval, just make the two violinists pop out in front of you. By closing your eyes, you canget fooled of having real instruments in front of you.
While I felt that the DI was a match for the Jkeny’s modified Hiface on the soundstage and frequency extension department, it felt though that, in my system, the modified Hiface had the edge in low level details. My personal guess is that the battery powerhelps throwing a blacker background and increase the perception of low level details.This is the only area where the battery powered Hiface has a clear and distinct edge over the DI.
Transients seemed subjectively faster on the DI than on the Battery power Hiface, which might also explain why I felt that the DI was being slightly more dynamic that the modified Hiface.
I believe that it has perhaps to do with the fact that John Kenny, the one who modified my Hiface aimed for avery “analog” sounding device while Kingwa, the designer of the DI aimed for arelatively neutral transport.
More importantly, the increase in transient speed doesn’t come at the expense of decay of sounds. In fact, I was quite surprised to notice that the decay of sounds was actually longer on the Digital Interface (versus the battery powered Hiface).
My guess is that Jkeny’s Hiface was tweaked to sound “analog” through smoothing the rough edges of the stock unit while the Digital Interface sounds “analog” by sounding more realistic and truer to life (instead of rounding off the edges).Or it could also be that there is a "synergy" between the Digital Interface and the DAC that I am using (which is from the same maker).
The attacks are sharp and the decays seem to hang longer in the air than any other USB converter I have heard to date. This extended decay is very different from the haze the Teralink X2 is plagued with which adds the same euphonic coloration and warmth over all material that goes through it.
While it was pretty obvious from the start that the DI had excellent macro-dynamics, it took me a few days of listening to discover all the micro-dynamics subtleties it is capable of. Playing a well recorded voice or solo instruments, I got drawn into the music like never before. On my ALO recabled Beyer T1s, every tonal inflexion or micro-dynamic shift of a voice or instrument was mesmerizing.
Since I got my T1s a few months ago, I had been pretty impressed with their macro-dynamic capabilities but would have never guessed they were capable of such micro-dynamics subtleties.
Summary onthe DI:
Quick summary on the DI:
- Excellent low level resolution (better than all my converters save for Jkeny’s Hiface)
- Excellent top to bottom coherency and very realistic timber: instruments and voices seem to be made from one piece
- Excellent frequency extension at the extremes (similar to Jkeny’s Hiface)
- Excellent soundstage and imaging capability (similar to Jkeny’s Hiface)
- Excellent transient and dynamic capabilities (the best I have heard from a usb converter)
- Excellent upsampling algorithm (better than many software based upsamplers)
- No support for 88.2K frequencies
- A little soft sounding at 44.1K (it can be overcome by upsampling to 96K)
- The performance of the external PS is affected by many external factors (power filter, usb cable...)
Overall, I would say that the DI performed at a similar level as the battery powered Hiface. While the modified Hiface was more upfront with the details, the Digital Interface had a clearer edge in top to bottom coherency as well as transient speed and preservation of the natural decay of complex (non-amplified) instruments.
The differences between them are small enough that it will probably come down to personal performance and system synergy. Both units are a few steps up from entry level converters such as the Musiland / Teralink-X2.
In my system, and to my ears, I could sump up by saying that the DI simply made beautiful music and didn’t call attention to itself. There are probably other converters out there that can beat it in soundstage size,resolution, dynamics... but the good thing with the DI is that when you listen to it (alone), it is very hard to pinpoint areas of weakness.
While it probably won’t replace a $1000 Empirical Audio usb transport, it will probably satisfy any person looking for their first good quality USB transport. My main and only complaint, as I stated earlier, is that the DI (like the Teralink-X2, Bravo,Stello U2...) cannot handle 88.2K, which might not be such a big deal for most people because of the material that is available is either at 44.1 (CDs), 48(movies) or 96 (most High Resolution music).
Conclusion of Part 2:
While I subjectively preferred the DI in my particular system, I believe that Jkeny’s Hiface MK1 is overall superior because its performance is more consistent from one system to another. The only thing you have to worry about when getting Jkeny’s Hiface isgetting a suitable battery charger and any functioning USB extender. With the DI, small tweaks (external power supply, power filter...) can have a rather large effect on the sound. So in that way, there is no guarantee that you will get the same results with the DI as reported here: they might be better or worse depending on your particular system and associated components.
If we are purely looking at the performance, Jkeny’s Hiface has the upper hand: I tried it with various DACs from different makers, using different computers (laptops and workstations) and the result has always been the same: excellent. It also allows for 88.2K and 176.4/192K playback.
However, if we factor in the ease of use, extra functionalities (upsampler, recloker) and build quality, the DI has the upper hand. So overall, it is a tie in my opinion, and it might come down to personal preference and what you are ready to settle for.
Last year, when I was recommending the Hiface to head-fiers, it was a very novel and breakthrough device and it was also cheaper than what it is selling for today (I paid 82.5 euros for it at the time). Today, there are many 24/96+ usb converters popping out everywhere.
In my personal opinion, the stock Hiface lost its value proposition, and I would rather recommend the cheaper Teralink-X2 for entry level set-ups, the Digital Interface (without power supply) for mid level set-ups.
At a slightly higher price both the A-gd Digital Interface (with the external power supply) and Jkeny’s modified Hiface seem to represent excellent price quality ratio. I can personally live with either one and I can’t say the same with other converters. For those willing to spend even more, there are very interesting USB converters such as the Audiophilleo, Empirical Audio converters, Weiss INT202... to name a few.