Thanks to Maxvla, I've recently been testing out the Matrix X-Sabre and Mini-I (DAC/amp w/ dual AD1955s). Much like I compared the Gungnir to the Hex, I spent a good portion of last night comparing the X-Sabre to the Hex. Over the past couple weeks, I also spent a good deal of time testing out different upsampling methods on the Hex with HQPlayer. I ran both subjective listening tests and measurements, both to see how well the filters improved the Hex purely from an objective standpoint and to see what sort of ringing each method brought to the table by testing square waves. In this post and thread, I'll be covering my thoughts on the Matrix products, comparisons to the Hex, and other DAC-related thoughts that stemmed from all this.

First, I want to make it clear that my experience with DACs and other equipment is relatively limited compared to many others here. It's no secret that I listen to a lot of heavy music, mostly progressive metal (i.e. Opeth, BTBAM, etc.), and I also have a preference for a more laid-back sound that focuses on overall tone/timbre and transients than a perfect frequency response or resolution abilities.

X-Sabre and Hex Comparison

With the Hex and Gungnir, I felt the two shared more similar characteristics than not. The differences between the Hex and X-Sabre were more immediately noticeable. In both comparisons, I wanted to mention that I did not use software upsampling on the Hex (all 16/44.1). The first thing that stuck out to me was the way the X-Sabre presented upper mids and treble. It was noticeably better than the Hex at resolving certain details, most apparent to me in how the sound of sliding fingers on the fret board of an acoustic guitar stood out. The small details were more overlooked on the Hex, though I would not necessarily say it lacked detail. On occasion, light cymbals in the background sounded more distinct and clear overall on the X-Sabre, though still present on the Hex. In these regards, the X-Sabre had a tight, clean, well-defined sound.

However, the texture quality seemed a bit rough on the X-Sabre, and it overall had a very slightly lean characteristic. I can see why many would like this, though, as it did give a nice sense of detail. However, I found it at times to be a touch too unnatural sounding. For example, Åkerfeldt does a lot of finger-picking acoustic work in Opeth, and I know he leans towards a smooth, liquid sound and style of play. The X-Sabre, while I did mention its ability to resolve guitar work well, almost made it sound as thought he was not finger-picking. In other words, the slightly lean, rough, maybe even a tiny bit sharp quality (slightly metallic?) of the X-Sabre somewhat undid that smooth, liquid sound and play style and seemed to place a pick in between his fingers. However, I can't say I found it too bothersome or fatiguing, and at times I did appreciate the extra resolution abilities of the X-Sabre over the Hex. In most cases, I preferred the trade off of seemingly better tone/timbre and body than better resolving, less rolled treble. I would often find myself accidentally listening longer on the Hex than the X-Sabre, though, so it is slightly fatiguing in comparison to the Hex. The Hex was easier for me to get "lost in," and I felt a stronger emotional connection to the music with it. This comes down to personal preference.

In terms of dynamics and bass response, it was harder to tell. I found them comparable in overall dynamics and might give a slight edge on bass response to the Hex, though there were times I thought the X-Sabre came through clearer in the bass. I believe this was album/song/genre dependent. Generally, the Hex sounded a tad more powerful and had a greater sense of body. In terms of soundstage, etc., I thought the X-Sabre had a greater sense of space to it. But, I don't think that's much of a surprise, as it seems NOS "colla pses" much of that and gives a more intimate sound, if you will. I like to think of it as more focused, but I know others disagree with me there. For a lot of my music, I think that more intimate sound works well. Oh, and I found vocals on the Hex to sound more natural and lifelike with a better sense of presence than on the X-Sabre. Drums in particular tended to have a better sense of body on the Hex as well, which I preferred.

Things got more difficult or even weirder to compare in areas like high-pitched, lightly played cymbals. The X-Sabre did a better job making those aspects noticeable and detailed, but the Hex did a better job with a sense of attack and quick transients despite somehow making most other aspects of the cymbal sound mushier or more laid-back. It sort of comes through as two different forms of detail, clarity, resolution ability, etc. And that ties into my other, wider thoughts.

Other Thoughts - Transients and Ringing (Take with grain of salt, YMMV, I'm not an expert)

After trying out a lot of upsampling filters on the Hex, which really can make a large difference in how it sounds, and comparing the Hex against the Gungnir and X-Sabre, I am beginning to think the ringing that comes with oversampling/upsampling filters can play an audible role in music. Or, at least, more of a role than some claim. When comparing software upsamping filters on the Hex, there was a slight, slight difference between methods without ringing at all and methods with/without pre-ringing (but with post-ringing), even if all other measured aspects/improvements were similar. Make no mistake, the most audible difference with software upsampling came from the boosted treble and lowered distortion/artifacts (i.e. less laid-back and mushy, better FR and resolution abilities, change of tone/timbre). However, using upsampling with ringing on certain music, I thought I perceived a slight degradation of transient abilities. Let me try to explain...

One thing I noticed with the Hex vs. Gungnir was that the Hex seemed to do a better job with quick attacks and micro-black spaces in between notes. I found the same to be true against the X-Sabre. It is most noticeable with drums, especially in fast, complex metal passages. Simply put, I thought the Hex did a better job with transients, in that they were clearer and better defined (better attack). I'm not sure I'm using the term "transient(s)" properly, but hopefully I make sense regardless. I could achieve a very similar effect by using various software upsampling methods on the Hex. Some methods were even a bit lean/strident. The best way I can describe it is a very slight haze or echo to attacks/transients that, in certain situations, was noticeable (again, quick, complex passages with lots of fast drumming seemed to highlight this). While these upsampling methods did help with the upper frequency response, gave less of a laid-back/mushy sound, and helped with the sense of space and soundstage, there was sometimes an ever-so-slight dull or lack of focus to otherwise powerful, clean attacks. Again, the difference is nowhere near as noticeable as what you positively gain elsewhere, and it's very hard to describe or put your finger on it. But there were audible differences between similar upsampling methods that had different ringing characteristics. In a sense, both frequency/distortion and transient abilities are a form of "detail," though in different regards (frequency domain vs. time domain, both are important to music).

When I tested upsampling methods that did not have any pre or post-ringing, I kind of got the best of both worlds. Despite being the least recommended upsampling filter in HQPlayer's manual, I liked the polynomial-2 method on the Hex. It overall boosts the treble response (it does have a sharper roll-off, but the roll-off starts much later than without upsampling) and reduces distortion/artifacts much like other upsampling methods did. This helped reduce that laid-back, slightly mushy quality to the upper registers, kept the more intimate soundstage and smaller sense of openness (or, IMO, greater focus), and yet kept the strong transient abilities without making it sound lean or strident. The poly-1 method brought some improvements, but it had even more of a treble roll off than the Hex did without upsampling, so it was mixed.

Now, I could be very wrong about all of this, and/or my ears and brain could have been deceiving me. But, personally, I think oversampling methods and ringing might play a larger role in music playback than some claim. I know some DAC manufacturers, like Ayre, have taken a closer look at this without just going straight to NOS-mode. In particular, I found their white paper on the topic to be interesting: http://www.ayre.com/white_papers/Ayre_MP_White_Paper.pdf

Like all other things, I think whether one values the frequency domain or the time domain more, it all comes down to personal preference. I think given that the benefits of most upsampling/oversampling methods to the frequency domain are likely more noticeable than the downsides to the time domain, it makes more sense to focus on the frequency domain for most people.  (This strongly reminds me of how I prefer smooth, fast motion on computer monitors over absolute colour accuracy. A 120Hz monitor with 2D LightBoost enabled, which means super fast, smooth motion and nearly eliminated pixel ghosting, is glorious to me compared to a slower IPS display with perfect colour and contrast accuracy. Most people care less about smooth, fast motion or simply are less sensitive to it than I...it can literally be the difference of feeling sick or not while playing a video game for me.)

Back to the X-Sabre, Conclusions, Upcoming, Etc.

Overall, I found a lot to like about the X-Sabre. It wasn't exactly to my tastes, but I could see it being a really good purchase for many people. I'd be happy to live with it if I had to, but I'd rather take the Gungnir for its smoothness if I couldn't use the Hex. In particular, it was nice comparing a DAC that resolves well to the Hex, though it did help me further solidify my preferences when it comes to DACs. I just wish the X-Sabre had a smoother treble characteristic, but that might just be the sabre chip. I do like the simplicity of the unit and its build quality. It is solid and heavy enough to likely be used as a murder tool, despite its compact size.

Later, I will spend some time with the Mini-I and come back with thoughts. I did spend a bit of time testing it with my HE-500 as a DAC and amp, and my initial impressions were pretty positive. It seems like a fairly nice combo unit for the price! I believe I used the word "nifty" a few times when trying it out.