Discussion in 'Digital: DACs, USB converters, decrapifiers' started by Maxx134, Jul 22, 2018.
Or mid-range town. Might be too much of a good thing. Those ATCs are sure nice though.
While I am sure that more bit depth needs more stable temps, I really think the longer warmup times are due to internal volume and extra-metal that needs to come up to temp.
Modi MB being the smallest has the shortest warmup time while the Yggdrasil is the longest for the opposite reason.
Check the quotes I posted from Mike in this post.
He basically said the same thing, I was just paraphrasing.
I wouldn't put all the balls in the jitter court. USB usually measures with slightly less jitter than any of my SPDIF CD transports. But the CD transports sound more typical of less jitter, less smeared, less hazy, more focused than even the best USB. Go figure.
We need more data. The thing is, I think all bets are off once jitter gets below a certain point. Similar to distortion where it seems that all bets are off once we get below -80db. The isn't 2004 anymore. All the jitter measurements I've seen from recent stuff is very very good and far better than was typical 15 years ago.
I'd rather this community not chase that final picosecond. Too easy to lose sight.
Easy. Triangulate with other people who have heard Convert-2, DA11, Solaris and compared to the Schiit DACs. There should be more impressions as these DACs are making the rounds.
I trust people in this community. Especially the folks on the $$$ DAC loaners.
Interesting that you categorized the Convert-2 as lacking a black background compared to the others. I think we hear the same thing but put it in different terms.
The Convert-2 I felt had a blacker canvas than Solaris (this could partially be the Solaris being much better at picking up low level information). However the Convert-2 sounded as if it had a light blanket or veil covering over the sounds. The Solaris while being a bit more dark gray, did not have such a blanket.
The difference is the quality and blackness of the canvas which the sounds are rendered upon (blackground) and the extent of veil over the sounds (clarity). Their net effect is very similar and sometimes hard to distinguish from each other. Sometimes I lop them into one category as "blackness / clarity".
It’s true: I was presuming the differences between sources should be attributed to jitter. There could also be noise on unshielded lines, or bad USB implementation, or who knows...
USB is poop compared to a good SPDIF source. I don’t care what measurements say.
Any DAC that sounds better with USB to me is just...Well, mind boggling. How do you fuck that up unless you have a truly revolutionary USB implementation? (Read: basically no product being sold on the market currently.)
The new Metrum chips are legit. Black background, good dynamics, natural timbre. Piano is a good test case. I have a piano in every swampy bodily crevice on my body, so I know how they sound IRL. Downsides are you gotta pay thousands to really hear the new chips as they’re meant to be heard and you still get the NOS sound, albeit as good as I’ve heard.
Anyway, sounds like another person has discovered that USB sucks. Nice to see a cheap SPDIF source that does so well and lets others see the light!
As a hardcore cd user and after extensive testing with various dac's and cd transports
i have come to the conclusion that for cd spinning the best quality for me was from a dedicated cd player.
So i have sold everything and bought my self a cd player just like the good old times of my youth!
@Hands I appreciate the agreement that Pavane is as good as NOS currently gets, and that it has accurate timbre (no small feat). I don’t understand the view that NOS sound is a “downside.” What, in your view, are the weaknesses of NOS?
The treble roll off is audible, for one. But it’s mild and something you can synergize with. Still, it’s kind of like having an LCD monitor hooked up via VGA and just slightly not calibrated right, so you get a slightly soft image. OS can sometimes sound like the opposite, ie too much sharpness.
Certain elements, like if you were to isolate just the head on a drum being hit, might sound a bit more nebulous than compared to OS. Cymbals sound a bit less metallic, in a bad way. I think it’s a mix of the high frequency roll off mixed with the characteristic lack of a digital filter. The latter, no filter, has some upsides and clear downsides. Unique trait nonetheless.
On the other hand, I find NOS to make individual elements sound more full bodied and 3D compared to OS. So, you may hear the deep resonance of a drum more on NOS whereas OS will focus on the head. NOS may be coloration or incorrect representation of truth in this case. A downside to this is layering is more likely to blend together on NOS even if stage depth is good.
NOS seems to have good rhythmic qualities. Generally good micro and macro dynamics. Stops and starts on a dime. None of that bullshit beat you upside the head fake dynamic shit. We’re talking true groove here.
Vocals in particular sound more natural to me on NOS. And given vocals are probably the most “real” thing out there and something we all have intimate experience with, I find that easiest to connect with when done well.
Ultimately, though, NOS tends to excel in the mids, but has less strengths in the high end. (Piano and vocals do well here in particular.) This leads to obvious side effects like a less sharp and focused sound, less macro details, etc. OTOH you could argue OS will never sound totally natural either. I see compromises either way.
I nominate @Hands’ response as one of the most useful posts in DAC Talk!
I fully agree these are the differences I hear. Given the gentle treble, I always expected NOS to have poor transients. But just the opposite was true with Adagio: as you say, incredible ability to stop/start on a dime.
The comibation of very natural timbre and very fast and accurate transients leads me to assume that the treble of NOS may actually be more accurate than OS treble—even though there is the perception of hearing more detail in the OS treble, I wonder how much of that “detail” is artifacts or emphasis of the reconstruction.
I would probably just add from my side to the above that OS DACs are more prone to listening fatigue. There is something slightly 'metallic', 'intense' or 'strident' about pretty much all OS filters I have heard so far.
Upfront soundstage position, lack of depth and layering of the instruments on the stage, lack of a feel of sound propagation from a location on the stage toward the listener. NOS does have less of a flat cutout effect compared to OS with individual instruments, but I attribute this to the close localization without depth cues (something far out in t he distance is going to appear more flat). Also, NOS has a lack of space and gives few cues to the size of the venue from proper reverb timing and rendering.
I don't totally agree with the vocals part. There are NOS DACs that don't do vocals the "NOS" way like the Holo and then there are ASRC-NOS DACs (MD RDAC) that do it to a large extent. There are some OS UltraAnalog and PCM1704 based DACs that achieve this "vocals" thing. Although an NOS UA or PCM1704 might do this "vocals" things more. I like this vocal coloration, but something inside me tells me it's a result of low levels of harmonic distortion across the board. I would cite Metrum Quad measurements from www.changstar.com as the Quad did the "vocals" things to an extreme.
As far as things starting or stopping on a dime, I don't know if I totally agree with this either. I think this is more of an architecture or power supply thing rather than NOS thing. None of the current lower-end and first-generation Metrum DACs do this. The MOS16 I have in the house is punchy, but more blurry than the Beef A2 DAC.
As for the high-end and fatigue, it's the filters. Do an OS linear phase filter with a small passband and bigass stopband to mimic similar slow-knee roll-off as an analog LPF for 44.1kHz and the effect will be similar. It's just that nobody does this from an engineering POV. We are trading frequency response linearity with ripple magnitude and duration. There is debate whether the ripple is audible. I think it is.
NOS emulation knee on OS
That might be the case with most NOS. for which I cannot speak—I have only heard Adagio—but for Adagio I’d say it has the most depth and 3dimensional stage of any of the DACs I recently tested.
@purr1n I should have clarified that I was largely talking about top tier NOS DACs. Mostly Pavane/Adagio. It’s truly frustrating that means you’re looking at probably spending $3K+ to get the real deal. Anything below that is, at best, only going to capture some of what I mentioned and screw up other areas or, at worst, only provide faint glimpses into what could be while never amounting to anything great. (Though I guess a true worse case would be the new AGD shit in NOS mode.)
Poor measurements, linearity, etc apply in some cases but don’t always correlate with NOS traits. A good, NOS TDA1541 DAC will measure very well and still hold some of these traits, whether it’s coloration or not. And the Pavane does fairly well too when measured. Leagues better than the first gen stuff, yet still retain that sort of full bodied, real sounding vocal performance, for example. And I’m talking about the core elements here, not the overblown versions (like from Quad) that may actually correlate to measurements (high distortion, high jitter).
The Holo Spring looked great on paper but was like a weird contrast boosted, nuance obliterated version of NOS. Oil painting to the max or something. Maybe it’s an architecture or discrete ladder DAC thing, assuming it’s not broken like AGD.
It’s tough to figure out. Same for determining how an OS DAC will sound based on measurements.
I’m inclined to agree there may be more to elements like rippling than we’d believe.
On the start/stop thing, it could be that Metrum always has really clean DACs (power, noise floor, etc). They also use a minimal output stage and seem to choose good OPTs for balanced to SE conversion on the highest end models. (I am, after all, mostly familiar with Metrum when it comes to NOS, so this may not apply to other brands.)
But even then, I hear the same thing when dropping to a 2x slow roll off OS filter. Maybe it’s spacial information being lost. Could be perceived as faster decay or loss of plankton. Could be seen as less noise, haze. Hard to say.
I am developing more of an appreciation for softer OS implementations. I like the balance of a good 2-4X filter. Maybe it’s the subtle effect of aging a few years making me not gravitate towards NOS so much. Maybe better amps.
The Convert-2 of NOS?
I'd probably go TOTL Metrum if TTs were not my primary source. Additional considerations for your DAC search: Berkeley Alpha 2 and Bricasti M1. Figured if you are comparing TOTL NOS, but may as well compare to TOTL sigma-delta oversampling designs.
What CD transports did you end up with?
I do notice better sound when playing a CD in my PC vs playing lossless files, but I think that could be just the higher quality files from the CD.
I am not sure. Have not heard the Convert 2 myself, but given what others have said about it...maybe?
I'd go more for TTs were it not for the music I listen to (modern prog rock/metal, which while can be mixed/mastered relatively well, still not suited for vinyl) and nervosa. The slightest IGD, noise and pops, pitch and tempo variations, and all the tweaking variables that go into it as such...I couldn't do it. Playing around with sources and DACs is one thing. TTs are just not good for my mental health.
I really like the M1SE. Spent a lot of time in the Bricasti room at a recent show, and have considered pick up a used one a few times. But digital source usage is just too infrequent, so meh. They're still north of 5k used.
BTW, Bezerkeley just released a new Alpha 3. I've seen a few of the previous models up for sale, but fuck me, the prices...
@purr1n how's the Lynx Hilo as a DAC?
Separate names with a comma.